Next Article in Journal
Suicide and Health Crisis in Extremadura: Impact of Confinement during COVID-19
Previous Article in Journal
Medial Sural Perforator “Nerve through Flap”: Anatomical Study and Clinical Application
 
 
Article
Peer-Review Record

Celecoxib in a Preclinical Model of Repetitive Mild Traumatic Brain Injury: Hippocampal Learning Deficits Persist with Inflammatory and Excitotoxic Neuroprotection

by Matthew Hiskens *, Rebecca Vella, Anthony Schneiders and Andrew Fenning
Reviewer 1: Anonymous
Reviewer 2:
Submission received: 9 February 2021 / Revised: 1 March 2021 / Accepted: 12 March 2021 / Published: 26 March 2021

Round 1

Reviewer 1 Report

please see attached pdf file

Comments for author File: Comments.pdf

Author Response

Reviewer 1

Major

1) Measurement of COX-2 expression

This is an astute comment and we thank Reviewer 1 for the opportunity to focus and tighten some of the key points of this paper while identifying another limitation of this work. Reviewer 1 is correct that our model has been adapted from the design of Briggs et al. (2016), and that neither this study nor that by Briggs investigated COX-2 expression temporally. We agree that it is a limitation that our study did not measure COX-2 expression, and we have added this to our discussion of limitations (see line 521). We made the decision not to include this measure for three reasons: 1) we felt previous studies had shown that head impact would induce COX-2 and celecoxib attenuate increased expression; 2) we wanted to focus on exploring the effects of celecoxib on the diverse secondary injury mechanisms; and 3) we included TNF as a downstream indicator of the cytokine profile that is induced by COX-2 following mTBI.

 

We chose not to directly measure COX-2 expression in the hippocampus and cortex following head impact as we felt that several previous reports had established this increase in COX-2. We thank Reviewer 1 for the study that was mentioned by Cernak and colleagues, which demonstrated the increase in COX-2 expression following injury and a subsequent improvement in cognitive function following nimesulide administration. We have included an explanation of this work in our manuscript (line 74). Nimesulide has again been investigated by Girgis and colleagues (2013) (http://0-dx-doi-org.brum.beds.ac.uk/10.1016/j.brainres.2012.10.049) in moderate TBI, where an increase in COX-2 was seen following injury that was inhibited by treatment (now discussed in line 355). Another study not included in our manuscript provided additional evidence of the increase in COX-2 expression as found by Strauss (2000), where COX-2 was elevated in the cortex and hippocampus of rats receiving a moderate TBI. In our paper we included the work of Dash and colleagues (2000), where we reported the role of celecoxib in neurological function, and we have revised this sentence to emphasize that increased COX-2 expression was mitigated in this treatment (line 351). In our paper we also discussed findings by Gopez and colleagues (2005), who showed that increased COX-2 expression was seen after TBI in the cortex and hippocampus that was attenuated by the COX-2 inhibitor DFU. These studies all assess COX-2 expression in the context of moderate TBI. Therefore, the most relevant study for us to discuss regarding the profile of COX-2 expression following injury, which we neglected in the submitted manuscript, is by Kelso and colleagues (2009) who examined the COX-2 response in mild TBI in COX-1 and COX-2 deficient mice compared to their matching wide-type controls. This work is now cited in line 524. While this study by Kelso involved a mild TBI induced by controlled cortical impact which is different from the repetitive mild injuries that we delivered, we felt that this work, and the others cited above, provided enough evidence that COX-2 was induced by TBI and attenuated by treatment.

  • While we recognize that celecoxib is a COX-2 inhibitor, instead of focusing on the COX-2 response to treatment, the concept of this paper was to explore the diverse secondary injury mechanisms that are induced by repetitive mTBI and the effects of celecoxib on these processes. In our introduction (line 49) we made the statement that ‘protective effects are not strictly resultant from reduced inflammation, as increased neuronal survival following COX-2 inhibition was not associated with decreased mRNA inflammation signaling (12).’ Therefore, we sought to look at other mechanisms – excitotoxicity, glial involvement, neuronal breakdown – to understand previously unknown actions of celecoxib. An important concept that we sought to explore was described in the context of epilepsy by Rawat et al. (2019) (https://0-doi-org.brum.beds.ac.uk/10.1186/s12974-019-1592-3) whereby the effects of COX-2 selective inhibitors appear to show a dual effect on neuroinflammation and neuroexcitability via separate and distinct mechanisms, with inhibition of PGE2 production quieting the calcium-glutamate axis of seizures with cellular inflammation and toxicity suppressed through reduced pro-inflammatory cytokine production. The work of Dehlaghi et al, (2019) (https://0-doi-org.brum.beds.ac.uk/3389/fneur.2019.00811) showed that additional effects of COX-2 inhibition were seen in modulating microglial cells and astrocytes. Therefore, we have expanded our introduction to include these studies and provide additional context for our expanded exploration of treatment effects.

In understanding the COX-2 related effects of celecoxib, instead of directly measuring COX-2, we looked at a downstream product of COX-2, the cytokine TNF. When COX-2 is induced by an inflammatory stimulus, there are three pathways of products that are produced: 1) prostaglandins and thromboxanes, 2) cytokines, and 3) proteases. Each of these groups of products further exacerbate inflammation. For this study we focused on the 2nd of these pathways involving cytokine production, and specifically measured the increased expression of TNF. In this way we measured TNF downstream as an inflammation marker that was partly an indicator of COX-2 type involvement. We recognize that COX-2 is not the only mediator that produces TNF in TBI, and therefore with additional resources it would have been valuable to assess other cytokines and COX-2 outputs from the other two pathways, and this has been added as a limitation in the paper (line 522). We have added clarification in our discussion on the TNF-mitigating treatment effects and how this might be related to the reduction in COX-2 expression.

 

To summarize, we agree with Reviewer 1 that knowledge of COX-2 expression in our rmTBI model is important, and while we regret that this could not be prioritized in the present study, future work should explore this relationship.

 

2) Number of animals per group

We thank the reviewer for picking up an ambiguity in our explanation of the number of animals used for analysis. Line 112 was relating to animal safety and was intended to mean that no animal was excluded from the study on account of excessive trauma. We have clarified this wording to reflect this meaning. We have also provided further clarification of the selection of animals for analysis in line 109.

Each treatment group did have an N=8. However, from the power analysis that we conducted from pilot data, we knew that we would be able to generate statistically significant data from fewer than this N=8, and a subset of these animals for was used for each analysis based on what was indicated from our power analysis. We included N=8 in our dosing and impacts due to the risk that we may have to exclude animals due to poor tolerance to these interventions, however as mentioned in the manuscript we had no adverse events. We could have used N=8 but due to resource limitations we chose to use the numbers that were indicated by our power analysis. Regarding the qRT-PCR data, it is generally agreed that as few as three biological replicates can be used https://0-doi-org.brum.beds.ac.uk/10.1105/tpc.108.061143

 

3) Neurological Severity Score

Figure 2B erroneously included neurological severity score (NSS), which was planned for the cognitive analysis but unable to be undertaken due to staffing. Figure 2B has been updated so that NSS is no longer included.

 

Minor

Line 27: A “period” has been included to fix this grammatical error.

Line 181: Units (mM) have been included for the aCSF concentration values.

Line 227: The ‘#’ is now appropriately included to Figure 3.

Reviewer 2 Report

Traumas (traumas-1123904)

 

Celecoxib in a preclinical model of repetitive mild traumatic brain injury: hippocampal learning deficits persist with inflammatory and excitotoxic neuroprotection

 

This manuscript presents a study of the effects of celecoxib (CEL), a selective COX-2 inhibitor, on learning (Morris water maze, MWM) and gene expression measures of inflammation and excitotoxic injury a repeated mild TBI mouse model.

The authors used C57BL/6J mice in a weight drop (25g from 1m) model of mild TBI, with the impact directed to produce a rotational force component. Mice were anestesized with isoflurane and received 15 injuries (IMP) over the course of 23 days. Celecoxib (50mg/kg/day) or saline (VEH) was given subcutaneously, starting 14 days prior to injury and continuing through all impacts. Two sets of 4 groups of 8 mice each received VEH+anesthesia, CEL+anesthesia, VEH+IMP, and CEL+IMP, with one set studied at an acute time point (1 - 2 days after end of TBIs) and the other set studied at a chronic time point (89 - 90 days after end of TBIs). After the 2 day MWM testing the animals were euthanized and cortex and hippocampus were dissected and mRNA extracted. RT-PCR for MAPT, GFAP, AIF1, GRIA1, TARDBP, TNF, NEFL, and GAPDH (control) was performed.

The authors report a delay in return of the righting reflex after anesthesia for VEH+IMP and CEL+IMP at impact 1-4, as well as some less pronouced delays for CEL+IMP at later impacts. At the acute time point mTBIs produced impaired performance in the MWM (trials 2-4) and CEL treatment produced greater impairment in trials 3-4. At the chronic time point deficits in the MWM were only noted for the CEL+IMP group at trials 3-4. Gene expression experiments demonstrate a variety of statisically significant differences, with a pattern of CEL treatment reducing the markers for excitotoxicity, inflammation, glial activation, and neuronal damage seen most clearly in the chronic hippocampus. In other time points/tissues, the pattern is less clear and CEL+anesthesia produces significant effects on its own.

The authors conclude that repeated mTBI in this model produces a learning deficit in the MWM (at the acute and chronic time points) and increase in expression of genes important for excitotoxicity, inflammation, glial activation, and neuronal damage. While treatment with CEL ameliorated the increases in gene expression, CEL treatment did not improve the latency to recovery of righting reflex and deficits in the MWM and at some points worsened the deficits.

Overall I feel the manuscript is of interest, but it would benefit from attention to several points.

 

1. If there are n=8 mice per group and no mice were excluded, why are n’s for the individual experiments less than 8?

 

2. What were the results of the testing for motor function deficits (line 172-173)?

 

3. Were there any differences in brain weights?

 

 

Minor points

 

1. The first sentence of the introduction has a grammatical error.

2. In Figure 3 I don’t see the # symbol used in the graph.

Author Response

Reviewer 2

1) If there are n=8 mice per group and no mice were excluded, why are n’s for the individual experiments less than 8?

We thank the reviewer for picking up an ambiguity in our explanation of the number of animals used for analysis. Line 112 was relating to animal safety and was intended to mean that no animal was excluded from the study on account of excessive trauma. We have clarified this wording to reflect this meaning. We have also provided further clarification of the selection of animals for analysis in line 109.

Each treatment group did have an N=8. However, from the power analysis that we conducted from pilot data, we knew that we would be able to generate statistically significant data from fewer than this N=8, and a subset of these animals for was used for each analysis based on what was indicated from our power analysis. We included N=8 in our dosing and impacts due to the risk that we may have to exclude animals due to poor tolerance to these interventions, however as mentioned in the manuscript we had no adverse events. We could have used N=8 but due to resource limitations we chose to use the numbers that were indicated by our power analysis. Regarding the qRT-PCR data, it is generally agreed that as few as three biological replicates can be used https://0-doi-org.brum.beds.ac.uk/10.1105/tpc.108.061143

 

2) What are the results of the testing for motor function deficits (line 172-173)?

Motor function was assessed via swim performance and coordination. Swim speed was not different between injured (0.25 +/- 0.01 m/s) and uninjured status (0.27 +/- 0.01 m/s, p = 0.19). Limb movement and coordination was assessed by video analysis of swim performance and displayed no abnormalities for any group.

 

3) Were there any differences in brain weights?

There were no statistically significant differences in brain weight for any group (p = 0.062) across both the acute or chronic groups. These weights were taken immediately following sacrifice. A statement on this has been added to the results section, line 180. The data is as follows, however this data has not been included in the manuscript.

 

Acute                            mean (+/- SD)

VEH + CON               0.430g (0.018)

CEL + CON                0.438g (0.017)

VEH + IMP                0.465g (0.026)

CEL + IMP                 0.445g (0.010)

 

Chronic

VEH + CON               0.463g (0.005)

CEL + CON                0.458g (0.020)

VEH + IMP                0.448g (0.005)

CEL + IMP                 0.448g (0.013)

 

Minor

The grammatical error in first sentence of the introduction has been corrected.

Figure 3 has been corrected so that the # is now appropriately included.

Round 2

Reviewer 1 Report

All of my concerns have been addressed

Back to TopTop