Latest / EHA Unplugged / Hereditary Hematological Malignancies (HHMs): A New Era of Discovery in MDS and AML (EHA2024)

Transcript

Hereditary Hematological Malignancies (HHMs): A New Era of Discovery in MDS and AML (EHA2024) · EHA Unplugged

0:06
In this episode of EHA Unplugged, we will revisit a
0:09
spotlight talk from EHA 2024. Doctor Kieran Tawana from
0:14
Addenbrooks Hospital in Cambridge dives into the topic
0:17
of hereditary hematological malignancies.
0:19
So the subject of my talk will be hereditary hematological
0:22
malignancies, a new era of discovery in myelodysplasia and
0:26
AML. So in brief, this talk will
0:29
cover the discovery of the hereditary hematological
0:32
malignancy syndromes, how we can recognise these in clinical
0:36
practice, some key features about DDX 41 variants, which are
0:40
now the most prevalent germline variant recognised worldwide, as
0:44
well as updates on the key transcription factor syndromes,
0:47
Rungs 1, GATA 2 and CBP Alpha. I'll then go on to talk about
0:52
disease evolution within families, in particular, the
0:54
important molecular insights we can gain from these case studies
0:58
and how we can integrate testing for germline variants into
1:02
clinical practice with this amassing information.
1:06
So hereditary hematological malignancy syndromes are
1:09
syndromes where two or more cases of cytopenia and
1:14
hematological malignancy occur within a family, and this is
1:17
typically myelodysplasia and acute myeloid leukemia.
1:21
They're thought to arise from pathogenic germline variants in
1:24
genes typically regulating hematoporesis, and they give
1:27
rise to an autosomal dominant predisposition in the majority
1:31
of cases. However, it's important to note
1:34
that there is variable penetrance, with some syndromes
1:37
associated with only 10% penetrance, whereas others have
1:40
a higher risk of developing disease with over 80 or 90%
1:44
penetrance. The latency of disease onset can
1:47
differ significantly as well, with cases described as young as
1:51
in infancy, while others develop disease over the age of 80
1:55
years. As the germline landscape has
1:57
evolved over the last decade or so, these cases are now thought
2:00
to account for five to 15% of all myelodysplasia and AML.
2:06
It's essential we recognize these syndromes because it
2:09
enables us to perform accurate risk assessments in patients to
2:13
personalized treatment, in particular modifying toxicities
2:17
from modalities such as radiotherapy and to initiate
2:20
cascade testing of family members, particularly where
2:23
we're considering related donor stem cell transplantation.
2:28
We must also be mindful that these patients are at risk of
2:31
secondary cancers and surveillance is essential.
2:33
The germline field really began in 1999 with the with the
2:39
identification of germline rung swan variants in six families
2:43
presenting with thrombocytopenia and myelodysplasia AML.
2:47
Since then, a number of genes have been described in the
2:50
landscape and they can broadly be categorized according to The
2:54
Who 2016 classification. So that is syndromes which
2:58
predispose to myeloid malignancies alone including the
3:01
CBP alpha variants and DDX 41 variants, those associated with
3:06
platelet dysfunction in particular RUNX 1 and CARD D-26
3:11
and ETV 6 and those which are associated with other organ
3:15
dysfunction including GATA 2 deficiency and the
3:18
telomeuropathies. I'd just like to mention several
3:21
of the syndromes which have been discovered in recent years and
3:24
this includes ERCC 6L2 variants, which is an autosomal recessive
3:28
disorder giving rise to early onset myelodysplasia and high
3:32
risk AML. Quite often the patients will
3:35
have a prodromal period of cytopenias or bone marrow
3:38
failure and it's essential to recognize them during this early
3:42
stage as predisposition to leukemia is often associated
3:45
with adverse outcomes. DHX 34 variants were first
3:51
described by Doctor Rhea Maichen ET al in 2020 and this is a
3:54
nonsense mediated decay gene which gave rise to the onset of
4:00
cytopenias and myeloid malignancy in several families
4:04
with ages ranging from 2:00 to over 20 years.
4:10
Germline variants in ERG were recently described as a late
4:13
breaking abstract ASH and this is a known oncogene.
4:16
Interestingly, individuals with this syndrome harboured typical
4:20
crossover features between RUNX 1 and GATA 2 deficiency
4:23
presenting with thrombocytopenia,
4:25
thrombocytopenia, myeloid malignancy as well as
4:28
lymphedema. So the majority of patients with
4:33
myelodysplasia and AML will have next generation sequencing
4:36
performed at diagnosis and quite typically with the germline
4:41
variants. We see that the variant allele
4:43
frequency remains stable throughout the disease course,
4:46
irrespective of disease or remission status.
4:50
There are important exceptions to this rule and this includes
4:53
the Sam, D9 and D9L syndromes. In these cases, the germline
4:57
variant can be typically lost through acquisition of monosomy
5:00
7 or deletion of chromosome 7 Q, the location of Sam, D9 and D9L.
5:07
Alternatively, patients may undergo uniparental diatomy
5:10
offence of of the wild type allyl leading to loss of the
5:15
germline allyl and this actually leads to recovery of bone marrow
5:18
function. Whilst the majority of germline
5:22
variants are inherited, de Novo events can occur in particularly
5:26
post psychotic events and and these can lead to mosaicism with
5:31
again typically low variant allele frequencies and this has
5:34
been described in the Leif Raumaini syndrome.
5:38
So how can we recognise these disorders in clinical practice?
5:41
Well, it's important to maintain high index of clinical suspicion
5:45
and be aware of the variable signs and symptoms.
5:48
There can be key morphological features including severe
5:52
megakaryocytic dysplasia, particularly with the
5:55
thrombocytopenia syndromes as well as gatitude deficiency.
5:59
And it's important to ask us a personal and family history
6:02
which may only be positive in about 40% of cases.
6:07
However, symptoms such as cytopenias, myelodysplasia,
6:10
acute leukemias, bleeding propensity should all raise a
6:13
red flag. The majority of patients with
6:17
germline variants present at a young age, typically less than
6:21
50 years. We've got A2 cases peaking in
6:24
the first or second decade. The majority of cases with CBP
6:28
alpha variants will present before the age of 40.
6:31
And the average age of onset for Rungs 1 variants is in the late
6:34
30s, although there is a peak slightly later in life.
6:38
With the bone marrow failure syndromes, we see that SAN, D9
6:41
and D9L are really restricted to younger age groups, and the
6:45
telomeuropathies can occur slightly later.
6:48
However, there's one syndrome that stands apart from the rest
6:51
and that is DDX 41, where patients typically have a peak
6:55
age of onset beyond 65. South DDX 41 variants
7:02
represented a novel molecular entity back in 2015 when they
7:06
were first described. This is a gene encoding an RNA
7:10
helicase, but it also has an important role in immune
7:13
function, in particular in the STING pathway, which mediates
7:15
responses to intracellular pathogens.
7:18
The majority of patients presented with myelodysplasia
7:21
and AML, with a median age of onset of 65.
7:25
Typically, patients had a hypocellular marrow with
7:27
erythroid dysplasia, and the majority of cases had a normal
7:31
carrier type. Interestingly, somatic DDX 41
7:35
variants were acquired in over 50% of cases, highlighting how
7:40
important bialelic disruption was for disease onset.
7:45
Since then there have been multiple studies mapping the
7:48
landscape of both germline and somatic DDX 41 variants and what
7:52
we see in the germline setting is that there is often several
7:56
key hotspots and these define founder mutations.
8:00
Somatic DDX 41 variants can occur throughout the gene but
8:03
typically target the R 525 codon.
8:06
So while we know these variants occur in the in the general
8:09
population, a recent study sought to assess the prevalence
8:14
of germline DDX 41 variants and they studied over 450,000 adults
8:21
of European and non European descent through UK Biobank data.
8:26
They found that germline variants were detected in one in
8:29
129 cases. However when criteria of
8:32
pathogenicity were used the the figure actually dropped to 1 in
8:37
430. This is really quite a high
8:41
figure for an autosomal dominant predisposition gene.
8:45
However, DDS 41 variants only give rise to about two to 4% of
8:50
all MD's and AML, so studies have focused on looking at which
8:55
factors are really associated with the onset of malignancy.
8:59
We know that truncating variants are thought to have a higher
9:02
risk of malignancy onset compared to MIS sense and start
9:06
loss mutations. There's also a high
9:09
preponderance for males. In the original study by
9:12
Pulpaserta Tau, 95% of cases with MD's and AML were male.
9:17
Subsequent studies have have suggested that the figure is
9:20
closer to a a ratio of three to one.
9:25
Interestingly, with GDX 41, this is a late onset malignancy.
9:28
However, clonal hematopoiesis is not observed in these cases, and
9:33
the key driver of disease is actually a somatic DDX 41
9:37
variant, which is now thought to arise in about 80% of cases.
9:41
Additional mutations do occur in CUX 1 and genus, also in RUNX 1,
9:46
ASXL 1 and TP 53, but there's no clearly defined molecular
9:50
signature. With the exception of this
9:51
somatic variant, multiple studies have elucidated the
9:56
functions of DDX 41, which include mRNA splicing, STING
10:01
activation as well as RNA processing.
10:04
However, we're yet to answer the key key question, which is why
10:07
are patients presenting so late? So in contrast to DDX 41, the
10:15
the transcription factor syndromes have more clearly
10:18
defined profiles. We can see here with Runks
10:22
Gemline RUNKS 1 variants the penetrance approaches 50% with
10:25
the majority of cases developing myelodysplasia and AML.
10:31
There is also an important but small number of cases presenting
10:33
with lymphoid malignancy including TALLBALL and non
10:37
Hodgkin's lymphoma. We now know that the majority of
10:42
RUNX 1 carriers do develop clonal hematopoiesis which
10:45
increases with age. The variants identified are TET
10:49
2B core and DNMT 3A and disease transformation is associated
10:54
with the acquisition of a somatic RUNX 1 variant
10:58
associated with other cytogenetic changes, PHF 6
11:02
variants and B core variants with GATA 2.
11:07
The penetrance is over 80% and the majority of cases develop
11:11
myelodysplasia. About 10 to 15% of cases will
11:14
present with AML. We know that clonal matter
11:18
process also occurs with GATA 2 carriers.
11:20
However, it's thought to be less common than with rungs one and
11:23
the dominant feature is usually asymptomatic carriage of an AS
11:27
XL1 variant. Disease transformation is
11:31
associated with the acquisition of monosomy 7, deletion of seven
11:34
Q and interestingly this derivative unbalanced
11:38
translocation between chromosome 1Q and seven Q, which appears to
11:42
be enriched in the GATA 2 deficiency syndromes.
11:49
Germline CBP alpha variants also have a very high penetrance with
11:54
which increases with age with over 80% of cases developing AML
11:59
de Novo. I'll go on to talk about this
12:03
syndrome now. This germline CBP Alpha variants
12:07
were first identified back in 2004 when two siblings presented
12:13
to Saint Bartholomew's Hospital back in 2003.
12:16
They were aged 18 and 30 and presented within a matter of
12:19
weeks with AML M2. On further questioning, their
12:24
father had also developed AML at the age of 10, and all three
12:28
individuals shared a germline and terminal CBP Alpha variant.
12:35
In the years that followed further CVP alpha, the pedigrees
12:38
were described and this prompted us to commence a collaborative
12:42
study to characterize these families in more detail.
12:46
We found that the germline mutations were all n-terminal
12:50
frameshift variants, which preceded the internal P30 start
12:54
codon. All tumors also acquired a
12:58
somatic CVP alpha variant which was located in the C terminal,
13:04
particularly targeting this lysine 313 residue and these
13:08
were typically in frame insertions or deletions.
13:14
When we mapped out the clinical events across this cohort, there
13:17
was a striking observation and that was that individuals had
13:21
multiple episodes of disease, often spanning a time course of
13:25
decades. The cumulative incidence of
13:28
relapse was 56% at 10 years and relapses often occurred late
13:32
with a median time to relapse of 27 months, but as late as 167
13:38
months. So this prompted us to study the
13:42
relapse samples and what we found was that the somatic C
13:46
terminal mutation actually differed at relapse in five out
13:50
of six cases and this phenomenon was observed as early as 13
13:54
months following the diagnostic presentation or as late as 14
13:59
years. So further further sequencing of
14:02
these pedicures using whole exome sequencing and deep
14:06
molecular sequencing identified that none of these C terminal
14:10
variants were actually present in the diagnostic sample,
14:14
confirming that that they were unique leukemic episodes.
14:20
The overall survival of familial CVP alpha AML was favourable
14:24
compared to even sporadic double mutant AML which is thought to
14:28
be associated with good outcomes and this was particularly
14:32
surprising considering that there are higher number of
14:34
relapses. So when we assessed the relapse
14:38
post survival, we found the answer to this.
14:41
The median relapse post survival was eight years compared to 16
14:45
months in sporadic subtypes, showing that relapse episodes
14:50
actually retain chemo sensitivity with favourable
14:53
outcomes. So in summary, this represents a
14:58
unique pattern of disease evolution which is thought to
15:00
arise due to the high penetrance of n-terminal germline variance.
15:04
Relapses are caused by new leukemic episodes which often
15:07
occur late, and we see favourable responses to second
15:11
line therapies and outcomes superior to even sporadic, say,
15:15
CBP alpha double mutant AML. However, the story continues.
15:21
In the years that followed, germline C terminal CBP alpha
15:25
variants were actually described, but these were
15:27
different to the variants which accompanied the N-terminal
15:31
frameshift lesions. They're instead missense or
15:34
frame shift lesions that were located in the leucine zipper
15:38
domain. And interestingly they're
15:41
associated with a far reduced penetrance with over with only
15:45
about 40 to 50% of carriers developing disease.
15:49
Disease onset occurred younger, younger on in life as expected
15:54
and the outcomes are still waiting to be fully
15:57
characterized but suggest that they are inferior to the germ
16:01
line and terminal variance. So whilst it's important to look
16:06
at larger family studies, we also gain important insights by
16:10
looking at case studies. And this was a particular family
16:15
that was referred to us from Hungary where the mother was an
16:18
asymptomatic carrier of a nonsense Runks 1 variant.
16:22
She had four daughters who all carried the same variant and,
16:26
interestingly, they all harboured the Jack 246 one
16:30
haplotype, which is found in 25 to 40% of European individuals
16:34
and predisposes to the risk of myeloproliferative disorders
16:38
harbouring the Jack 2V-617-F variant.
16:44
We performed whole exome sequencing across all four
16:49
children because unfortunately three daughters developed AML
16:54
and the third sibling had MD's. In the siblings that developed
16:58
AML, we found acquisition of an SH2V3 variant associated with
17:03
uniparental dysomey increasing the mutant Allo burden.
17:07
The second twin developed a Jack 2V-617-F mutation, again with
17:11
uniparental Dysomey, and the fourth sibling who presented
17:15
later again at the age of five years of age had acquired a Jack
17:20
2V-617-F mutation. So we had three siblings with
17:24
AML all presenting at the age of 5 and acquiring Jack stat
17:28
variants in the context of a 46 one Hapner type.
17:32
So this really suggests that pathogenic germline variants and
17:36
host genetic features can interact to influence somatic
17:40
mutation acquisition. So this phenomenon of convergent
17:45
evolution isn't restricted to Rungs 1 pedigrees.
17:47
We've also seen it occur in GATA 2 pedigrees and this was one
17:50
such pedigree where we studied 3 cousins. 2 cousins presented
17:55
with high risk MD's and AML and unfortunately died with
18:00
complications post stem cell transplantation, but they both
18:04
harboured this somatic AS XL1 variant.
18:07
The third cousin presented over 10 years later, again with the
18:11
same AS XL1 variant but at a lower variant allele frequency
18:15
and she only had mild neutropenia.
18:19
Clinical surveillance was performed, performed for this
18:21
individual over a period of time, and this longitudinal
18:27
monitoring actually allowed one of our colleagues, Dr. Al
18:29
Sarahi, to actually study the mechanisms of penetrance in this
18:34
case. In the earlier years, she
18:36
observed monoallelic expression of the mutant GATA 2 allele,
18:40
which coincided with more severe monocytopenia and neutropenia as
18:46
time progressed. Interestingly, the counts
18:48
recovered and this coincided with recovery of wild type GATA
18:52
2 expression. She tried to investigate the
18:57
reasons for this and actually found that this was due to
19:00
increased methylation of the wild type allele promoter
19:03
leading to selective expression of the mutant allele.
19:08
Interestingly, with this higher mutant expression, there was a
19:11
higher higher AS XL1 variant allele frequency, which
19:15
correlated with disease severity.
19:19
So this really suggests that clonal evolution and disease
19:21
manifestations are governed by mutant allele expression.
19:26
So as we've learnt a great deal over the years with germline
19:31
variants, it's important to think about how we can optimize
19:34
testing of patients in clinical pathways.
19:38
Now, when patients present with high risk MD's and AML, they're
19:42
often considered for stem cell transplantation.
19:44
Within three to four months of presentation, they will undergo
19:48
tumor testing and then if suspicious variants are
19:51
identified, they will need confirmatory germline testing
19:55
followed by related donor testing if they find a suitable
19:59
match. The key issues here are the the
20:03
culture of skin fibroblasts can take up to four to six weeks and
20:06
leads to significant delays in testing.
20:09
So we would therefore advocate that if a hereditary
20:12
hematological malignancy syndrome is suspected, a skin
20:16
biopsy should be taken at diagnosis, preferably with the
20:19
diagnostic trefine biopsy, and sent the culture of skin
20:23
fibroblasts to prevent delays later on.
20:26
We're fortunate in the UK to have access to whole genome
20:29
sequencing, which can be performed upfront.
20:33
However, where germline results are unavailable and time is
20:38
limited, we would advocate parallel testing of both the
20:41
patient and the related donor to exclude an underlying germline
20:47
predisposition syndrome. Now we have a number of
20:52
modalities to test myelodysplasia and AML.
20:55
However, we do need to ensure that our assays are
20:58
comprehensively set up to identify all germline variants
21:04
and this includes non coding lesions, synonymous variants and
21:09
copy number variants. We know that non coding lesions
21:12
can account for up to 10% of GATA 2 cases and not and are not
21:16
often not included in typical next generation sequencing
21:19
assays. Synonymous variants can be very
21:22
difficult to interpret without further studies and copy number
21:27
variants again require optimization of either SNIP
21:30
arrays or FISH panels to ensure we detect these either single
21:36
exon or partial allelic deletions.
21:40
It's as we identify more and more genes, it's essential that
21:42
we update our gene panels accordingly.
21:45
However, even with thorough comprehensive genetic testing,
21:50
we know that 40% of families remain uncharacterized.
21:58
So in conclusion, the assessment of germline predisposition is
22:02
now an integral component of genetic testing in myeloid
22:05
malignancies. Inter and intra familial disease
22:09
evolution highlights new genetic associations and our and
22:14
advances our understanding of disease.
22:17
It's a dynamic field and requires optimization and
22:20
standardization of testing. And as clinicians, we must
22:24
maintain clinical vigilance and consider collaborations with
22:27
research groups to advance further developments and
22:31
discovery. Thank you.