Animal models/60s/extra info Flashcards
Construct validity
different definitions -
essentially - how well does the underlying manipulation reflect pathology
- how much the model is depresssion
Pierro - aetiological basis
Face validity
how well does the animal model mimic phenotype and pathological changes seen in disease - degree of similarity between the model and the disorder.
how much the model mimics depression
Pierro - recapitulation of external features of NDDs
Predictive validity
how well does the model respond to treatments for given disease, but some authors also use it in a broader sense, including a models aptitude to predict some specific markers of disease
How will structure my essay? (animal models)
- Discussing the insights they can give use into disease pathology
- The aspects of disease pathology they can and cannot model
- The aspects of behavioural outcome they can and cannot model
- And finally considering alternatives such as IPSCs
Seok et al. (2013)
- conducted a large-scale study that highlights how poorly mouse models of inflammation reflect the human condition
- This involved analysing the gene expression profiles of leukocytes in mouse models and human subjects that were elicited by inflammatory challenges (including trauma, sepsis and endotoxemia).
- authors found that the expression profiles amongst humans across these diverse conditions were similar, suggesting a common pathway in humans independent of the initiating stress
- However, the expression profile in the corresponding mice models differed significantly from each other, and from the human data
- In fact, the correlations of the gene changes between humans and mice was equivalent to what would be expected by random chance
- This may explain why so many of the clinical trials for these disorders fail, and emphasises the need for more human-representative models, such as the DRAG mouse model (humanised mice - express human HLA-II)
- There were however caveats to this work, for example they did only investigate one strain of mice, different microbiotal compositions in species could have influenced the profiles, and humans received medical care that wasn’t accounted for
Issues with chronic social defeat
- Tsankova 2006 – increase in BDNF III and IV – but generally cant model chronic social defeat in female animals, although depression is twice as common in female humans
Issues w FST
- Immobility could be from shock of being dropped into the water, or could instead demonstrate learning or habituation (would therefore be a positive behavioural adaptation)
Issues with TST
- Can be weeks before a noticeable effect is observed in patients who take ADs – TST measures one AD dose for 5-6 mins
Issues w MWM
- – to evaluate impared cognitive function – have to remember where the platform is
- Has been shown that differences in performance are dependent on differences in thigmotaxis between rodents – tendency to stay close to the edges
- Differences in spatial memory may only be one causal factor
issues with SA
- Might just be animals’ curiosity
Discuss pre vs post dictive validitiy
- Predictive validity – ability to predict response to new drugs
- Postdictive validity – sensitivity to existing drugs
- The vast majority of studies have relied on a limited subset of tests
Handley 1993
-
1993 for example and show different findings depending on setup for SSRIs, but are consistent for benzodiazepines.
- EPM always works/is consistent for benzos
- For 5-HT targeting drugs it depends on light levels
- This is concerning given that several SSRIs are approved for various anxiety disorders
- Shows that many preclinical models lack postdictive validity
- Cav – they only tested 2 light intensities – would have been better to test a range
- Mouse defence test battery
- Model defence-related disorders such as anxiety or panic
- Incudes several measures of mouse defensiveness in response to the test
Stuart 2013
- developed affective bias task
- animal is given two independent learning experiences and then asked ‘which do you prefer?’ during a preference test (in this case association between digging cue and fixed food value) – mouse digs in 2 places, will get reward in either
- Following learning, present animals with both previously rewarded substrates during pref test - resulting choice bias score is then calculated to determine if the treatment has induced a positive or negative bias
- idea behind the task is that the animal will re-activate its memory for the reward associated with each substrate and then bias its responding based on the relative value it attributes to each experience
- Replicated affective bias changes seen in humans from acute SSRIs - suggesting translational and predictive validity
- Pharmacological interventions producing negative affective bias in humans also did so in rats.
- Good predictive and translational validity - however further work is required to see if this task has construct validity
Discuss benzos and issues
· Benzodiazepines - discovered in 60s, still prescribed today for anxiety disorders
§ GABA potentiators
§ Serendipitous discovery
· Sedative, motor incoordination, dizziness, anterograde amnesia. Real worry is risk of tolerance, dependence and OD. Usually cause serious OD in the presence of other CNS depressants, particularly alcohol - respiratory depression
Fauholt-Jepsen et al 2015
Our digital biomarkers are generated by sensors, voice and speech recognition, and other mechanisms when we use digital tools such as smartphones or smartwatches
- Faurholt-Jepsen et al. (2015) showed that smartphone data can be used as an electronic biomarker to gain insight into illness activity in bipolar disorder
- Their study involved 61 patients aged 18-60 diagnosed with bipolar disorder (BD)
- The authors used a software for smartphones (called the MONARCA I system) that collects automatically generated objective data, as well as self-monitored data on illness - activity
- Results:
- They found over time significant positive correlations between BD scores (for both mania and depression) on different classifications and the number of incoming/outgoing calls per day, the duration of calls, as well as the number of outgoing text messages per day
- They also reported significant negative correlations between self-monitored data (i.e. mood and activity) and depressive scores, and significant positive correlations between self-monitored data (i.e. mood and activity) and mania scores
- Moreover, using the automatically-generated objective data, they were able to discriminate between affective states
- There were some caveats to this study, including the small sample size of 33 patients (although patients were followed up for a median 310 days), there was also an larger proportion of participants with type II bipolar disorder compared to type I, and patients received different types, doses and combinations of psychopharmacological treatments throughout.
- May serve as a biomarker for illness activity for each patient
- Could be used to subcategorise disease and generate new approaches to behavioural therapy
- Ethical challenges here – too invasive/data leakage?
Musselman 2001
conducted a double-blind study on 40 patients with malignant melanoma who received IFN-α therapy; half the group was given paroxetine (SSRI) and the other half was given placebo. Development of depressive symptoms was significantly reduced in the paroxetine group compared to control, showing that the IFN-α behavioural syndrome is sensitive to regular antidepressants.
CONVERGE consortium 2015
performed low-coverage whole-genome sequencing on over 5000 Chinese women with MDD, selected to minimise phenotypic heterogeneity. They identified two significant risk loci at the SIRT1 and LHPP regions; this was replicated in an independent cohort of Chinese women. The function of LHPP is poorly defined, but SIRT1 is known to play an important role in mitochondrial function; therefore, mitochondrial abnormalities may be linked to depression risk.
Harmer et al. (2004)
Reducing the negative perceptual bias
conducted a study where healthy volunteers were randomised in a double blind fashion to placebo or SSRI. At the end of the study measures of emotional processing were tested using subliminal presentation of fearful and happy faces. SSRI decreased recognition of negative facial expressions and significantly increased positive bias in facial expression perception compared to placebo. This was hypothesised to be connected to effects on the amygdala, as this structure plays a key role in the primary processing of negatively valenced input. Therefore, they assessed amygdala responses using BOLD fMRI. They found that response of the amygdala to fearful faces was significantly reduced in volunteers receiving citalopram compared to those receiving placebo.
i.Yilmaz 2021
Overexpressed human C4A in mice
IHC and behavioural testing
Increased microglial engulfment and reduced synapse density, shown both in the visual system and the PFC
In the PFC this was slow and cumulative – maybe this is why scz develops later on
These mice also had increased social withdrawal and deficits in working memory
Mice without the C4 had normal cortical synapses
Maybe C4 could be a therapeutic target?
Yoshimizu et al 2015
i.used iPSCs to look at functional phenotype associated with an SNP identified in GWAS
Generated iPSCs with the allele and without
Found higher expression of CACNA1C in those carrying, confirmed by higher VGCC in those carrying by whole cell recording
Although iPSCs not same as in vivo
Lintunen 2021
longitudinal prospective study showed patients on calcium channel blockers had reduction in scz rehopsitalisation (although concomitant AP was allowed)
Karunakaran 2019 (Nature)
was all computational analysis
Identified the protein-protein interactions (interactome) of scz associated genes.
Looked at which drugs target the interactome and whether they show anticorrelation with the transcriptomic signature seen in the disease (transcriptomics)
Identified a shortlist of 12 potential drugs (top 3 were cromoglicic acid, acetazolamide and cinnarizine).
No actual screening to see if such drugs alter scz phenotypes – could have done animal model for each repurposed drugs
- conducted a large-scale study that highlights how poorly mouse models of inflammation reflect the human condition
- This involved analysing the gene expression profiles of leukocytes in mouse models and human subjects that were elicited by inflammatory challenges (including trauma, sepsis and endotoxemia).
- authors found that the expression profiles amongst humans across these diverse conditions were similar, suggesting a common pathway in humans independent of the initiating stress
- However, the expression profile in the corresponding mice models differed significantly from each other, and from the human data
- In fact, the correlations of the gene changes between humans and mice was equivalent to what would be expected by random chance
- This may explain why so many of the clinical trials for these disorders fail, and emphasises the need for more human-representative models, such as the DRAG mouse model (humanised mice - express human HLA-II)
- There were however caveats to this work, for example they did only investigate one strain of mice, different microbiotal compositions in species could have influenced the profiles, and humans received medical care that wasn’t accounted for
Seok et al. (2013)
-
1993 for example and show different findings depending on setup for SSRIs, but are consistent for benzodiazepines.
- EPM always works/is consistent for benzos
- For 5-HT targeting drugs it depends on light levels
- This is concerning given that several SSRIs are approved for various anxiety disorders
- Shows that many preclinical models lack postdictive validity
- Cav – they only tested 2 light intensities – would have been better to test a range
Handley 1993
- developed affective bias task
- animal is given two independent learning experiences and then asked ‘which do you prefer?’ during a preference test (in this case association between digging cue and fixed food value) – mouse digs in 2 places, will get reward in either
- Following learning, present animals with both previously rewarded substrates during pref test - resulting choice bias score is then calculated to determine if the treatment has induced a positive or negative bias
- idea behind the task is that the animal will re-activate its memory for the reward associated with each substrate and then bias its responding based on the relative value it attributes to each experience
- Replicated affective bias changes seen in humans from acute SSRIs - suggesting translational and predictive validity
- Pharmacological interventions producing negative affective bias in humans also did so in rats.
- Good predictive and translational validity - however further work is required to see if this task has construct validity
Stuart 2013
Our digital biomarkers are generated by sensors, voice and speech recognition, and other mechanisms when we use digital tools such as smartphones or smartwatches
- Faurholt-Jepsen et al. (2015) showed that smartphone data can be used as an electronic biomarker to gain insight into illness activity in bipolar disorder
- Their study involved 61 patients aged 18-60 diagnosed with bipolar disorder (BD)
- The authors used a software for smartphones (called the MONARCA I system) that collects automatically generated objective data, as well as self-monitored data on illness - activity
- Results:
- They found over time significant positive correlations between BD scores (for both mania and depression) on different classifications and the number of incoming/outgoing calls per day, the duration of calls, as well as the number of outgoing text messages per day
- They also reported significant negative correlations between self-monitored data (i.e. mood and activity) and depressive scores, and significant positive correlations between self-monitored data (i.e. mood and activity) and mania scores
- Moreover, using the automatically-generated objective data, they were able to discriminate between affective states
- There were some caveats to this study, including the small sample size of 33 patients (although patients were followed up for a median 310 days), there was also an larger proportion of participants with type II bipolar disorder compared to type I, and patients received different types, doses and combinations of psychopharmacological treatments throughout.
- May serve as a biomarker for illness activity for each patient
- Could be used to subcategorise disease and generate new approaches to behavioural therapy
- Ethical challenges here – too invasive/data leakage?
Fauholt-Jepsen et al 2015
conducted a double-blind study on 40 patients with malignant melanoma who received IFN-α therapy; half the group was given paroxetine (SSRI) and the other half was given placebo. Development of depressive symptoms was significantly reduced in the paroxetine group compared to control, showing that the IFN-α behavioural syndrome is sensitive to regular antidepressants.
Musselman 2001
performed low-coverage whole-genome sequencing on over 5000 Chinese women with MDD, selected to minimise phenotypic heterogeneity. They identified two significant risk loci at the SIRT1 and LHPP regions; this was replicated in an independent cohort of Chinese women. The function of LHPP is poorly defined, but SIRT1 is known to play an important role in mitochondrial function; therefore, mitochondrial abnormalities may be linked to depression risk.
CONVERGE consortium 2015
Reducing the negative perceptual bias
conducted a study where healthy volunteers were randomised in a double blind fashion to placebo or SSRI. At the end of the study measures of emotional processing were tested using subliminal presentation of fearful and happy faces. SSRI decreased recognition of negative facial expressions and significantly increased positive bias in facial expression perception compared to placebo. This was hypothesised to be connected to effects on the amygdala, as this structure plays a key role in the primary processing of negatively valenced input. Therefore, they assessed amygdala responses using BOLD fMRI. They found that response of the amygdala to fearful faces was significantly reduced in volunteers receiving citalopram compared to those receiving placebo.
Harmer et al. (2004)
Overexpressed human C4A in mice
IHC and behavioural testing
Increased microglial engulfment and reduced synapse density, shown both in the visual system and the PFC
In the PFC this was slow and cumulative – maybe this is why scz develops later on
These mice also had increased social withdrawal and deficits in working memory
Mice without the C4 had normal cortical synapses
Maybe C4 could be a therapeutic target?
i.Yilmaz 2021
i.used iPSCs to look at functional phenotype associated with an SNP identified in GWAS
Generated iPSCs with the allele and without
Found higher expression of CACNA1C in those carrying, confirmed by higher VGCC in those carrying by whole cell recording
Although iPSCs not same as in vivo
Yoshimizu et al 2015
longitudinal prospective study showed patients on calcium channel blockers had reduction in scz rehopsitalisation (although concomitant AP was allowed)
Lintunen 2021
was all computational analysis
Identified the protein-protein interactions (interactome) of scz associated genes.
Looked at which drugs target the interactome and whether they show anticorrelation with the transcriptomic signature seen in the disease (transcriptomics)
Identified a shortlist of 12 potential drugs (top 3 were cromoglicic acid, acetazolamide and cinnarizine).
No actual screening to see if such drugs alter scz phenotypes – could have done animal model for each repurposed drugs
Karunakaran 2019 (Nature)