Yale Psychiatry Grand Rounds: Lustman Awards 2024

June 14, 2024

Information

June 14, 2024

Lustman Awards

First Place: Robin Bonomi, MD, PhD: "Evidence of HDAC6 Suppression in PTSD"
Honorable Mention: Yang Jae Lee, MD: "Differences in Mental Illness Stigma by Disorder and Gender: Population-Based Vignette Randomized Experiment in Rural Uganda"
Honorable Mention: Sarah Abdallah, MD: "Characterizing Rare DNA Copy-Number Variants in Pediatric-Obsessive Compulsive Disorder"

ID11790

To CiteDCA Citation Guide

00:00Thanks, Kelly for that
00:02very nice introduction.
00:07Well, thank you everybody for being here.
00:08I'm excited to talk to you
00:10about my research today.
00:11I'm going to talk to you a bit about the
00:15histone deacetylase 6 and it's work in PTSD.
00:19We have some exciting translational
00:21evidence of suppression in PTSD
00:24for this enzyme we term HDAC 6.
00:28So you know PTSDI think most of us in
00:31psychiatry are familiar with this disorder,
00:33but it is anxiety,
00:35it's under the anxiety disorder
00:38diagnosis within our DSM 5.
00:40The treatments that we have currently
00:42that are FDA approved and most commonly
00:45used are primarily symptomatic treatments.
00:48And you know, with the leading treatment
00:50as starting treatment as our SSRIs,
00:52our selective serotonin reuptake inhibitors.
00:55But we know that we only have a 50 to
00:5860% response rate and and that's sort of
01:00on the best end with these treatments.
01:03And, you know,
01:04we have a variety of other symptomatic
01:05treatments we can go to as second,
01:07third, fourth lines to treat
01:09different parts of the disease,
01:11the nightmares, the insomnia, flashbacks.
01:14But we have a hard time really
01:17targeting the entire illness.
01:19Part of that is because this
01:21illness is so pleotropic and,
01:22and presents in so many different ways and,
01:25and at different times.
01:26It can present within, you know,
01:28six months after a trauma,
01:30it anytime after one month after the trauma,
01:32it is considered PTSD and,
01:35and no longer the acute stress diagnosis.
01:37But it can also, you know,
01:38we've seen Vietnam Veterans who
01:40present with symptoms of PTSD
01:4230-40 years after their trauma.
01:45And so how is that, how is that happening?
01:48And how,
01:48how is this all one disorder that
01:50we are trying to treat in one way?
01:52And, and what does that really mean?
01:54And,
01:55and how can we better understand this
01:57disorder and the pathophysiology of it
02:00such that we can target our treatments?
02:03And so I think, you know,
02:04the field as a whole is really starting
02:06to follow in the footsteps of,
02:07of the oncology research and of cancer.
02:09And we're,
02:10I think a little bit behind them
02:12in this precision medicine pursuit
02:13and this pursuit to,
02:15to find targeted treatments
02:17for individuals and,
02:18and where we can really help with that.
02:21I think maybe I'm biased,
02:23but I feel that H Dax and epigenetics,
02:25which I'll get into a bit more
02:27and utilizing neuroimaging,
02:28non invasive imaging modalities to be
02:31able to better understand the brain,
02:33allow us to,
02:35to really better understand these
02:37disorders on an individual basis.
02:39And so I think, you know,
02:41that allows us to hopefully down
02:42the road be able to stratify,
02:44stratify medications based on
02:47an individual's pathophysiology
02:48rather than trying to treat this,
02:51especially this pleiotropic
02:52disease as one and one only.
02:56So epigenetics and histone deacetylases,
02:58they're a class of 18 enzymes and they're,
03:01they're canonically known as
03:02epigenetic enzymes because of their
03:05role deacetylating or removing
03:06acetylomorides from the lysine
03:08tails on the histone core proteins.
03:11And so these tails are very lysine rich.
03:13They can be added an acetylm Y,
03:15they can be added or removed and that
03:17alters how closely the DNA is held
03:19and whether or not an open reading
03:21frame can be formed and therefore
03:24genomic transcription can occur.
03:25So the the H docs as they're known,
03:28it is 18 enzymes.
03:30However, within the H doc classes,
03:32there are four classes and they
03:34all function very differently.
03:36So I think H docs one through
03:38three are sort of very much this,
03:41this epigenetic role, but one that
03:42we are going to talk about today.
03:44H doc 6 is part of class 2B and it
03:48actually is more more predominant in
03:51the cytoplasm than in the nucleus.
03:52So it does shuttle between the cytoplasm and
03:55the nucleus of the cell and of the neuron,
03:57but in the brain,
03:58but it is actually more functional
04:00in the cytoplasm.
04:01So it's not quite an epigenetic enzyme in the
04:04in the way that we think about epigenetics.
04:07But it was originally found to be a
04:11microtubule deacetylase in its role
04:13of deacidylating the alpha tubulin at
04:15lysine 40 and in that it stabilized
04:18and destabilized alpha tubulin and
04:20part of the microtubule structure.
04:22And so that allows for both
04:24the growth cone formation,
04:25so cell,
04:27cell growth as well as cell some
04:30of cell movement through filopod
04:32formation and then also replication.
04:35As we know,
04:37mitophase and anaphase rely
04:39predominantly on the microtubules
04:42and then it also has a role in
04:45autophagosome degradation and and
04:47it was found then through some
04:49studies about 10 years ago to be
04:51abundantly expressed within the
04:52serotonergic neurons of the brain and
04:54so within the dorsal Raphae nucleus.
04:56They found this,
04:58this highly studied,
05:00highly cited study found that H
05:03.6 was very highly expressed in
05:05the dorsal Raphae nucleus and
05:06within the serotonergic neurons.
05:08And moreover,
05:09they found that if they knocked
05:11it out both constitutively with
05:13a a Cree driven deletion of H .6
05:15in mice as well as if they did a
05:18pharmacologic inhibition with H .6
05:20selective inhibitors that they could
05:22promote different and a less anxiety
05:25prone phenotype of these mice.
05:27So this led to this understanding
05:29that you know there may be something
05:32going on here with H .6 in emotional
05:34regulation and an anxiety regulation
05:37and they began to better understand
05:40that H .6 was needed for the role
05:43of the glucocorticoid receptor.
05:45And through that and through a
05:48few other studies that were done
05:50over the next couple of years,
05:51this role for HDAC 6IN deacidylating and
05:55therefore stabilizing or destabilizing
05:57the HSP 90 glucocorticoid receptor and
06:01FKBP 5 two complex was better understood.
06:04And So what actually under happens
06:07is the HDAC 6 removes this acetyl
06:10moiety from HSP 90,
06:11which is our heat shock protein that
06:13is a a commonly known chaperone protein
06:15for the glucocorticoid receptor.
06:17And that allows it to complex with
06:20the glucocorticoid receptor when
06:22a glucocorticoid such as cortisol
06:23binds and with FKBB 5.
06:25And that allows this translocation of
06:27the glucocorticoid receptor from the
06:29cytoplasm into the nucleus and then
06:32therefore the glucocorticoid response
06:33elements to be transcribed on the genome.
06:36And so the thought was and,
06:38and based on their deletion studies,
06:40they, the sort of the predominant
06:43understanding was that an increase in
06:45H Doc 6 therefore would lead to this
06:47increase in stress susceptibility
06:49because we felt that an increase
06:51in H Doc six would allow more of
06:54these GR ES to be expressed.
06:57And so to better understand that in in the
07:01in vivo in humans as well as in animals,
07:04there is this novel tracer that
07:06has been developed out of MGH and
07:08the Martinez imaging Center there.
07:11And they developed this tracer
07:13called Beveristat,
07:14which is a fluorine 18 labeled PET
07:17suitable radio lag and targeting HDAC 6.
07:20And so this tracer has had had gone
07:22into humans first in Belgium and
07:24then we redid quite a bit of the test
07:27retest studies here at Yale as well.
07:30And we decided to put this into to
07:32PTSD both in humans as well as in
07:34a rodent model to better understand
07:36how HDAC 6 and if we could is,
07:38is acting both in humans in PTSD.
07:42And so our original hypothesis was
07:44that the stress susceptibility or PTSD
07:47symptoms would be correlated with
07:48an increase in HDAC 6 based on these
07:51findings from these 2012 and 2015 papers.
07:54And we believed that,
07:56you know,
07:57this translational imaging would
07:58allow for this visualization
08:00and quantification of H .6.
08:01And so the rodent model that we chose to
08:03use was a single prolonged stress model.
08:06I used predominantly all male
08:08Sprite Dolly rats at this point.
08:10That's a,
08:10we do need to go into females of rats.
08:13I, I am not, that's not lost on me,
08:16but for the,
08:16for the proof of principle of this study,
08:19we did all males and we use the
08:21single prolonged stress model
08:22for a couple of reasons.
08:23One is that it's well validated in in rats,
08:27which we cannot use mice as well with
08:30PET imaging due to the size of the brain.
08:33The other is that I felt that it
08:36really capitulated a nice model
08:37of the pro anxiety and the hyper
08:40arousal symptoms of PTSD.
08:42And it also recapitulated this single
08:44stress event rather than some of the
08:48more prolonged stress paradigms that
08:51I think sometimes lead to more of a
08:54depressive phenotype rather than a
08:56a stress or anxiety and PTSD like phenotype.
08:59So this procedure involves restraining
09:01an animal for 120 minutes,
09:04putting them in a tank by themselves
09:06to swim for 20 minutes,
09:07letting them rest for 15 minutes.
09:09And then they are exposed to diethyl
09:11ether until they lose consciousness.
09:13At which point they are single housed
09:15and untouched with free access to
09:17food and water on a traditional 12
09:19hour light dark cycle for seven days.
09:21And it has been shown repeatedly in
09:23the literature that at this point,
09:25after seven days, they will demonstrate
09:28APTSD like phenotype or a hyper arousal,
09:30high anxiety like phenotype.
09:34So the first experiment that
09:36we did was really, you know,
09:37the, the typical experiment,
09:38I think where, you know,
09:40we took two groups of animals.
09:41We had one group that was stressed.
09:43We had one group that was controlled.
09:44They were handled both similarly
09:46leading up to the stress day.
09:48After the stress day, the,
09:49the control animals were handled,
09:51but no but continue to remain double housed,
09:54but we're not single housed after that.
09:56And the stressed animals were
09:58stressed and then single housed
10:00and not handled after that point.
10:02They then both underwent open field
10:04behavioral testing and H .6 pet on
10:06the same day and the following day
10:07they were sacrificed and their brains
10:09were removed for post mortem analysis.
10:13And you know,
10:15pretty much as we expected based
10:16on the literature,
10:17we did see changes consistent with
10:20anxiety phenotype and consistent with
10:22what was shown in the literature
10:24to be this PTSD like model in
10:27our behavioral testing.
10:29And then interestingly,
10:31when we did the H stack 6
10:33imaging with Beveristat,
10:34we found a decrease in the stressed
10:37animals in these key regions of
10:38the brain that we identified.
10:41So and we didn't see, you know,
10:43there was a slight decrease in
10:44the cerebellum in the whole brain,
10:46but not statistically significant,
10:47unlike the other key regions that
10:50are affected more under stress
10:52such as the amygdala,
10:53hippocampus,
10:54limbic cortex and the brain stem,
10:56notably where the dorsal Raphae
10:58nucleus and serotonergic neurons
10:59are predominantly located.
11:00And this was really surprising to us
11:02given our original hypothesis was
11:04that there would be an increase.
11:06And so because of this and because
11:08of the quantification with PET
11:10in an animal without an arterial
11:12input function to rely on,
11:13it's a bit more difficult.
11:16We decided to do it within subject analysis.
11:18And so I took rats and we
11:21I imaged them at baseline,
11:22we stressed them and then imaged them
11:24again to see what this would look like.
11:26And we actually found similarly
11:28the same thing. So we found both.
11:30We found a decrease in the striatum
11:33and amygdala, hippocampus,
11:34as well as a non significant decrease
11:37but trending in brain stem and limbic cortex.
11:39And when we stratified all of these
11:41rats by low anxiety and high anxiety,
11:43we found that there was a difference.
11:44There was a slightly more of a
11:47decrease in the high anxiety group
11:49as compared to the low anxiety
11:50in most of the regions,
11:52although it was not always
11:53statistically significant.
11:57So this led to this next question,
11:59which was, you know,
12:00are we catching these animals?
12:02Are we catching this too late?
12:04You know, maybe this is a,
12:05this is a compensatory decrease after
12:07an increase and maybe the the knockouts
12:10that were done before stress in these
12:13preclinical studies were knocking out
12:15H stack 6 before it increased and
12:17therefore they were seeing this response.
12:20And now we're seeing sort of the fallout,
12:21which is the a larger swing in the
12:24opposite direction because we really
12:25were trying to understand, you know,
12:28why are we seeing this decrease?
12:30So the next study was to do
12:32an acute stress study.
12:34And so we took the animals,
12:35we split them into two groups,
12:36again, did the same paradigm,
12:38stress versus control,
12:39but instead only waited 48 hours before
12:42doing behavioral testing and imaging.
12:44And as predicted based on the literature,
12:47at 48 hours,
12:48they do not show a significant
12:49difference in the anxiety phenotype.
12:51So they really don't look that much
12:53different than the controls in the way that
12:55they're interacting with their environment.
12:56However,
12:57they do show significant differences
12:59or very close to significant,
13:01I should say.
13:02They're they're not quite there,
13:03but in the acute stress time period.
13:07So as you can see,
13:09they show almost equivalent brain changes
13:11of H .6 to the PTSD animals at 48 hours.
13:15So this was really striking to
13:17us because it demonstrated to us
13:19that before the behavior changes,
13:21we are seeing these brain changes
13:23in H .6 that are preceding the
13:26actual behavior of PTSD like model.
13:28And we also did Western blocks which
13:31demonstrated that at two days post
13:33stress in the prefrontal cortex we
13:35were seeing a significant decrease
13:37in the glucocorticoid receptor.
13:39This was total cell,
13:41total tissue glucocorticoid
13:43receptor concentration,
13:44which I think is a little bit hard
13:45to to understand exactly what that
13:47means when we're talking about a
13:48translocation from the cytoplasm to
13:50the nucleus rather than up or down.
13:52But I think it was, you know,
13:53is it interesting finding to go along
13:55with this decrease in HDAC 6 as well.
14:00So to change over to humans,
14:03this is what we found in the animals.
14:04And now we were interested to see does the,
14:07do these findings hold in humans
14:09with PTSD as they did in rats?
14:11Because again,
14:12we don't know what the translation
14:14is necessarily from rat rats to humans.
14:16We know that the rodent models of
14:19PTSD are good, but not always great.
14:21And we know that the pathophysiology of
14:24PTSD in particular as a psycho pathology
14:26and as a psychiatric illness is,
14:28is not one that is well recapitulated in
14:31animal models consistently because of its
14:34difficulty to really get to the the human.
14:37I think the human nature of the
14:39flashbacks and the memories are
14:40very difficult to know that we're,
14:41we're actually putting that into an animal.
14:45And so moving into human work,
14:47we, we took a sub,
14:50took three groups of humans,
14:51healthy controls, trauma,
14:53exposed controls and PTSD.
14:55They were relatively well matched.
14:58There were some slight age differences,
14:59but otherwise relatively well matched
15:01groups with the exception notably of course,
15:04of their scores.
15:05So we looked at key PTSD scores that we use
15:10particularly in research such as the CAPS 5,
15:12the PCL 5 and then we did a number
15:14of others and addressed the HAM
15:16DMA to better understand their
15:18anxiety and depressive phenotype.
15:21And when we looked at this,
15:25we,
15:25So we imaged all of these individuals
15:27and found that we did see significant
15:30decreases in the PTSD individuals as
15:33compared to the healthy controls.
15:35But we also saw decrease in
15:38the trauma controls,
15:39most notably in the dorsal Raphae nucleus.
15:41And we have the dorsal Raphae
15:43nucleus we know is sort of a,
15:45it's a long,
15:47a long nuclei across the brain stem.
15:49And so this encompasses both the midbrain
15:53sort of locus and the ponds locus.
15:55And so we for the purposes of
15:56quantification of the imaging,
15:57we have them separated.
15:59But it was notable to us both
16:01that there was a decrease which
16:03was consistent with the animals,
16:05but also that that this was lower
16:08in the trauma controls and not just
16:12in the PTSD and that there was no
16:14difference between the trauma control
16:16and the PTSD significant or statistically.
16:18Sorry.
16:20So going back to our original hypothesis,
16:23why are we seeing,
16:24you know, this?
16:25We originally hypothesize we
16:26would see higher HDAC 6.
16:28We're now seeing lower HDAC 6.
16:31And what what does that mean?
16:32And so,
16:33you know,
16:34we had originally hypothesized
16:35that that that increased HDAC 6
16:38would relate to increased stress
16:40susceptibility and more translocation.
16:42But we also have this prior evidence from
16:45another excellent resident and physician,
16:48scientist Doctor Shivani Bhatt,
16:50and with her mentorship under
16:53Doctor Cosgrove that looked
16:55at some cortisol imaging.
16:57And so this was looking at an enzyme
16:59that produces cortisol in the brain
17:01and they found that this was increased
17:03in the brain in PTSD individuals
17:05as compared to trauma controls.
17:07So you know,
17:09there's a,
17:10I'm wondering now we're
17:11wondering if there's a
17:12chance that this is a compensatory
17:14response if there is increased cortisol
17:17centrally and that therefore the HDAC 6 is,
17:20is sort of trying to down regulate that
17:23response and decreasing the glucocorticoid
17:26receptor translocation into the nucleus.
17:29And so, you know, is that is that,
17:31is that a possibility?
17:33I think we, we don't know yet,
17:35but that's, that's sort of one
17:38response to this hypothesis.
17:40And, and you know,
17:41the difference between the trauma
17:43control and the PTSDI think one,
17:45one note that we're also interested in is,
17:48is could this lower H .6 be a
17:51trauma marker that's indicating
17:53this predisposition for PTSD?
17:55You know, we don't we,
17:56we know that many people who undergo
17:59multiple traumas are more likely after
18:01their first traumas or after childhood
18:03trauma to acquire PTSD later in life,
18:06either from that trauma or
18:07from subsequent traumas.
18:08And so, you know,
18:10is this evidence that this may be one of the
18:14changes that is predisposing them to PTSD?
18:16Certainly not all individuals.
18:17Some individuals,
18:18it is their first trauma and
18:20only trauma that triggers PTSD,
18:22of course.
18:23But what you know what,
18:25what's happening here?
18:25Why are we seeing this change both in
18:28the animals at this acute period and in
18:30the humans and these trauma controls,
18:32as well as in the PTSD individuals?
18:34And so I think that's part of an
18:37ongoing question that we need more,
18:40more studies to better answer.
18:41But this was an exciting
18:43first step to to understand
18:48in summary, you know,
18:49we contrary to the hypothesis,
18:51we now see lower levels of
18:54HDAC 6 both in trauma and in,
18:57in following trauma in
18:58both humans and in animals.
19:00These do we do consider these
19:02pretty robust findings given the
19:04reproducibility across multiple cohorts
19:06of animals within subject and across
19:09groups as well as within humans.
19:11And then, you know,
19:12this would imply that there's a
19:14decreased glucocorticoid receptor
19:16translocation into the nucleus and,
19:19and we have these hypothesis about
19:21possibly a compensatory mechanism
19:22due to increased central cortisol,
19:25but I think more studies are needed
19:27to better understand that full system.
19:31And with that, I'd like to thank
19:32you all for listening. And I'll,
19:34I think we have time for questions.
19:37So I can take any questions.
19:38And of course,
19:39thank both Doctor Cosgrove and Dr.
19:41Matt Dugante, who's sitting here as well,
19:43who's also been a,
19:44a foundational mentor to me and has really
19:47supervised all of the animal work and,
19:50and blended quite a bit of
19:51resources from his lab.
19:52So we thank him as well.
19:54Thank you.
19:58Thank you, Tom for that
20:00really nice introduction.
20:02So I'm I'm excited to be able
20:05to talk to all of you about my
20:07research trying to understand the
20:10role of rare genetic variants
20:12in the development of OCD.
20:19So I'll start just by describing
20:21some of what we already know
20:24about the genetic basis of OCD.
20:26So firstly, we know that it's
20:28moderately to highly heritable.
20:31So this means that the proportion
20:33of the variants in the trait,
20:35basically whether or not someone has OCD,
20:38is determined in part by genetic factors,
20:41and that heritability is higher for the
20:44childhood early onset form of the disorder.
20:47We also know that OCD is polygenic,
20:51so there's a contribution from
20:54estimated hundreds of genes,
20:57all that can contribute to OCD etiology,
21:00but we don't know what
21:02most of those genes are.
21:04And that's ultimately what our
21:06goal is from this research,
21:08is to identify more of those
21:10genes that could tell us about
21:13the underlying biology and could
21:15be potential novel drug targets.
21:18The
21:18other thing we know is that
21:20there's a contribution
21:21from many different types
21:23of genetic variations.
21:25So there's contribution from common
21:28variants that are relatively
21:30common in the human population.
21:32This is what we study with
21:35genome wide association studies.
21:36There's a contribution from rare
21:39inherited variants that are passed
21:42along through families and there's
21:45a contribution from Genovo variants,
21:47which our lab is particularly interested in.
21:52So Genovo genetic variants are not
21:56passed down from parent to child.
21:58They are spontaneous DNA mutations that
22:02arise from an error in DNA replication.
22:06And we all have a few of these,
22:08but they're relatively rare.
22:10And because they're so rare,
22:13they can be used to statistically implicate
22:16genes in the development of a disorder.
22:20And our lab was actually able to
22:23do this previously by looking at
22:26small de Novo variants in OCD.
22:30So by small,
22:31I mean point mutations and and small
22:34insertions or deletions of DNA sequence.
22:37And we found an increase in the rate
22:40of these small de Novo variants that
22:43were likely to damage the protein
22:46product of a gene in children
22:49with OCD compared to controls,
22:51which suggests that this type
22:53of variation was playing some
22:55role in development of OCD.
22:57And we were able to use these de Novo
23:00variants to implicate 2 risk genes that
23:04are thought to be associated with OCD.
23:10So we were interested in using a similar
23:13methodology and applying it to other
23:17types of de Novo genetic variants
23:20and other types of rare variation.
23:22So one thing we were interested in
23:24looking at was copy number variants.
23:26So these are large deletions or
23:30duplications of chunks of DNA sequence,
23:34so much larger changes than
23:36what we were looking at before.
23:38And these have been identified
23:41in patients with OCD,
23:43They've been studied before,
23:46but there's not yet a good evidence
23:50base showing their association
23:52with the development of OCD.
23:56So we opted to use our previous data set,
24:02whole exome sequencing data in
24:06trios with children who had OCD.
24:09So these were families where the
24:12parents were unaffected and the
24:15children have OCD and comparing
24:17this to unaffected control families
24:20with children who don't have OCD.
24:22So we performed quality control
24:26and used this data set to identify
24:31in silico copy number variants,
24:34classify them and do a comparison
24:36of the mutation rates.
24:41So we did quality control with PCA and
24:45made sure the cohorts were ethnically
24:47matched and these are the demographics of
24:50the samples that passed quality control.
24:57And what we found was that rare
25:01deletions were increased occurring
25:04at both an increased rates and a
25:08greater proportion of children with
25:11OCD compared to children without.
25:14And this effect was particularly
25:16marked for de Novo deletions.
25:23And looking at some of the more
25:26specific features of the de Novo CN BS
25:29that we identified in our OCD sample,
25:32a few things jump out.
25:33The 1st is that a lot of these individuals
25:37have Co occurring psychiatric disorder,
25:41most notably Tourette's, which we know
25:43goes along with OCD quite commonly.
25:45So we did a secondary analysis
25:48removing the individuals with the
25:50Tourette's diagnosis and found
25:52that the finding still holds.
25:54The other thing to note is that
25:57a lot most of the mutations,
26:00the copy number variants that we
26:04identified in individuals with OCD
26:06are predicted to be pathogenic
26:09or likely pathogenic,
26:10meaning they're disrupting
26:12or affecting gene products.
26:15And this is in comparison to the
26:18control samples where all the
26:20CN VS that we identified were
26:23not predicted to be pathogenic.
26:25So that further suggests that
26:27there may be some role for CN VS
26:30that are predicted to be damaging
26:32in the development of OCD.
26:38So what we can say from this finding
26:42is that de Novo rare, rare deletions,
26:46but primarily de Novo deletions
26:48are more common in children
26:51with OCD compared to controls.
26:53Most of the de Novo CN BS that
26:56we identified in our sample are
26:58predicted to be pathogenic.
27:00And this suggests that de Novo CN
27:03BS might play a role in the etiology
27:07of OCD and we could use this to
27:10help implicate new OCD risk genes.
27:13This is a really robust approach
27:16that at this point has identified
27:19100 some risk genes and autism,
27:22and we're hoping that it'll lead
27:24to the identification of more risk
27:27genes in OCD that could be potential
27:29drug targets someday and help us
27:31understand the biology of the disorder.
27:36So I'd like to end by thanking everyone
27:39who was involved in this research and
27:43in my clinical and research training,
27:46many of whom are are here in the audience.
27:48So thank you for your support and
27:51for funding sources for the award
27:54committee and my wonderful colleagues.
28:02Hello, my name is Jay.
28:03I'm a PG I3 resident.
28:05I'm sorry that I could
28:06not be there in person.
28:08I'm on my honeymoon now.
28:09So so I'll try to be available for
28:12questions after this presentation,
28:14but I'm sorry that I cannot be
28:16be there with you in person.
28:18So my presentations on mental health
28:21stigma by gender uncondition in
28:24rural Uganda and I'm excited to
28:26share with you our findings today.
28:29So a little bit of background.
28:31So mental illness are the leading
28:34cause of disability globally and 80%
28:36of those with mental illness reside
28:38in no and middle income countries
28:40as defined by the World Bank.
28:43But most of like the resources
28:44and most of the research done,
28:46mental illness is done in high income
28:49countries and a major factor in mental
28:52illnesses are mental illness stigma.
28:54So it's very prevalent globally and it,
28:58you know,
28:58it has many downstream effects
29:00such as levels of funding
29:01available for mental health care,
29:03treatment seeking behaviors and
29:06treatment outcomes as well.
29:10So in terms of characterizing mental
29:12illness stigma, as we all know,
29:14mental illnesses can be very,
29:16it can have very varying presentations.
29:19It can be schizophrenia,
29:20it can be anxiety,
29:21it can be alcohol use disorder.
29:23There's many different
29:25types of mental illnesses.
29:27So there have been studies in high income
29:30countries characterizing, you know,
29:32how people receive these different diverse,
29:35you know, mental illnesses.
29:37But there have not been so many studies
29:40in low and middle income countries
29:43characterizing the differences in how
29:45people perceive these different conditions.
29:48So in high income countries,
29:49the research shows that those whose
29:52symptoms align with gender based
29:54stereotypes are more susceptible
29:56to experiencing stigma.
29:58But, you know,
29:59those studies,
29:59to our knowledge,
30:01differentiate mental illness stigma by
30:03gender or condition in low income countries,
30:06to our knowledge.
30:06And that's, and by gender,
30:09you know, it means you know,
30:10the gender of like the person that's
30:13experiencing mental illness and the
30:14gender of the person that's that's
30:16evaluating or that's endorsing
30:18stigmatizing attitudes towards
30:19that person with mental illness.
30:21So,
30:22so both from both perspectives in
30:25understanding how people perceive
30:28these different mental illnesses
30:30are very helpful because then
30:33anti stigma interventions can
30:35be designed accordingly to that.
30:38So now I'm going to talk about
30:40methods and you know,
30:41I'm going to break this down step by step.
30:44So first, you know,
30:45we design some vignettes that describes
30:47a person with a mental illness.
30:49Then we randomly use random
30:51sampling methods to select the
30:52vignette to read a participant.
30:54And then we administered a survey with
30:58two with two with two parameters,
31:02one's measuring proximal personal stigma
31:04and another one measuring personal
31:08stigma about the person in the vignette.
31:12So for the vignette development,
31:15for developing these vignettes,
31:17we developed 4 vignettes illustrating
31:19alcohol use disorder, anxiety disorder,
31:22depression, and schizophrenia.
31:24And we created both a male and
31:26a female version of those,
31:28so you know,
31:31111 So each condition had two vignettes,
31:331 depicting depicting a man with
31:35a vignette and another depicting
31:36a woman with a vignette.
31:38In the vignettes,
31:39depression and schizophrenia were
31:41adapted from a study in Western
31:43Uganda and vignettes and alcohol use
31:46disorder and anxiety were developed
31:48utilizing the SM5 criteria and and
31:50having an intercultural team as I
31:53was involved as well as Ugandan
31:56collaborators as well to make sure
31:59that it is adaptable and under and
32:02understandable to the local context.
32:05So this is where we did the study.
32:07You can see Puyenda district is
32:09highlighted in purple on the map.
32:11This is by Lake Choga.
32:12It's a very rural district.
32:14It's got about four,
32:15a little bit over 400,000 people.
32:17It's also a very underserved area
32:20of Uganda which is a low income
32:23country and majority of population
32:25are subsistence farmers.
32:26There's 45% illiteracy rate above
32:2918 years old.
32:33So the study procedure,
32:34so we selected 20 villages and three
32:37sub counties in Vienna district.
32:40Each sub county has about 50,000 people.
32:42So we have a list of all the
32:44villages in these three sub
32:46counties and we utilize a number,
32:48a random number generator to to pick
32:50which villages we're going to do.
32:52So each village is numbered
32:54and we picked them randomly.
32:55And then once we arrived at the village,
32:58the community health workers have a
33:00registry of all households in the village.
33:03And so we again use the random number
33:06generator to select households.
33:08And once we selected a household,
33:10we again use a random number
33:12generator to select an adult,
33:14defined as 18 years or above
33:17in each household to interview.
33:19And then once we arrived at the household,
33:22we randomly selected the vignette.
33:24So there are 8 possible choices
33:27also with a random number generator
33:30for each participant.
33:32And this is a tool that we use
33:34to measure the level of stigma
33:36is adapted from a doctor Bob
33:39Rosenhach studies SO2 batteries.
33:41We're measuring a stigma and one
33:43is personal acceptance Scale,
33:45which measures proximal interpersonal stigma.
33:48So what I mean by proximal personal
33:51interpersonal stigma is, you know,
33:53what would that person do?
33:55So would that person be willing
33:57to have somebody with mental
33:59illness as a neighbor or not?
34:01And then we have another scale,
34:03which should be named a
34:04broad acceptance scale,
34:05which measures distal interpersonal stigma.
34:08So that's more questions like
34:09what should a society do about
34:11this person with mental illness?
34:16And so going into results.
34:18So these are some descriptive statistics.
34:20And you can see that we we got a pretty even
34:23distribution across all four conditions.
34:26And, and you know,
34:28we have more females and males, but you know,
34:32it's pretty evenly distributed otherwise.
34:34And this is the results of the
34:37personal Acceptance scale by condition,
34:40gender of the vignette and
34:41sex of the participant.
34:43So higher numbers means greater
34:45acceptance or less stigma.
34:46So this measure is personal acceptance scale.
34:49So so as you can see here,
34:52you know, in yellow in the in
34:54the bottom is schizophrenia.
34:56You know, it's got lower acceptance
34:58scores than let's say anxiety overall.
35:00And you know,
35:01you can also see some gender differences.
35:04So, so filled in circles means
35:07you know it's a female to female.
35:10So you know,
35:12both the person being depicted in
35:14the vignette is female and the person
35:16that is being read to is a female.
35:18And hollow circles mean that the
35:21person that's being detected is a
35:23female or being depicted as female
35:25and the person judging is a male.
35:27And, you know,
35:28likewise for and you can read
35:30the legends to to describe the,
35:33the the the rest.
35:36And this is a broad acceptance
35:37scale which had a, you know,
35:39somewhat similar story,
35:40although with less effect size
35:43than personal acceptance scale.
35:45So after accounting for multiple testing
35:48and model this specification via permutation,
35:51have you found significant differences in
35:53levels of stigma among the four conditions?
35:55For PAS, the P value is .00235
35:59and BAS the P value is .04757.
36:03So as you can see,
36:05the PAS was more had a higher degree
36:09of had a lower P value than the BAS.
36:12And then we did a pairwise
36:14secondary analysis and we found
36:16that acceptance differed in among
36:18anxiety and schizophrenia according
36:19to PAS and BAS significantly.
36:24So also we found some differences
36:27in acceptances across gender.
36:29So,
36:30so one thing that we so the PAS
36:35the varied across gender combinations
36:37for depression with the empirical P
36:40value of .017 and we can interpret this
36:42result as a reflection of personal
36:44acceptance being higher for women with
36:47depression than men with depression.
36:49If you refer back to our figure and
36:51other specific gender effects were
36:53not observed for any of the other
36:56mental disorders in the PAS or for
36:58any of the four disorders on the PAS.
37:00So discussion so significant findings.
37:03So we found significant differences in
37:04the levels of acceptance for depression,
37:07anxiety, alcohol use disorder,
37:08and schizophrenia after adjusting for
37:10the gender of the respondent and gender
37:12of the individual depicted in vignette.
37:14And schizophrenia was the least
37:16accepted in pairwise comparisons.
37:18After adjusting for multiple testing
37:20found that people with anxiety
37:21disorder appear to be more accepted
37:24than people with schizophrenia and
37:25women with depression were more
37:27accepted than men with depression.
37:29So compared to existing literature
37:31may mean high income countries.
37:33You know,
37:34our studies backed up the finding that
37:37schizophrenia has higher levels of stigma
37:39and depression and anxiety predictably.
37:41And but contrary to the existing literature
37:43that finds the mental illnesses that are
37:46more gender typical or more stigmatized,
37:48we found the opposite.
37:49Like for example,
37:50men with depression were were more
37:54stigmatized than women with depression.
37:57And you know,
37:58some other literature suggested
38:00substance use disorders are more
38:02stigmatized in mental illnesses.
38:03So it's also possible that pair wise
38:06comparisons between depression and
38:08alcohol use disorder and between
38:10depression and schizophrenia would
38:11have demonstrated significant
38:12differences with a larger sample without
38:15the Bloemfair or any correction.
38:17It did demonstrate significant differences.
38:20So strengths of the study were that the
38:22measurement of differential levels of
38:24stigma based on specific mental illnesses,
38:26adjusting for the gender of the
38:27respondent and the gender of the
38:29individual depicted in the vignette.
38:31And this hasn't really been done
38:33before in low income country setting.
38:35And given the importance of
38:37community and family support,
38:39especially in low and
38:40middle income countries,
38:42on outcomes for mental illnesses,
38:44the utilization of general population
38:48sample is particularly important
38:50in some weaknesses of the study.
38:51Or that translations can lead
38:53to subtle shifts in meaning,
38:55possibly affecting the nature
38:56of the responses elicited.
38:58And you know,
38:59it's also vulnerable to
39:00social desirability bias,
39:01like many surveys are so
39:04meaning of the study.
39:06Mental illness stigma is a major
39:07barrier to treatment and access.
39:09But most studies measuring mental
39:11illness stigma do not differentiate among
39:14different conditions and our study does.
39:16And this also in our differential
39:18findings also suggest that anti
39:20stigma efforts need to be disorder
39:23specific and gender specific.
39:24And we hope that the static and form
39:27the foundation for more targeted anti
39:29stigma approaches for mental illness.
39:31So there's some references and
39:34some acknowledgements below,
39:36and I'm happy to take any
39:39questions that people have now.