Transcript Personalized Medicine of Deafness

Personalized Medicine of Deafness
Marci M. Lesperance, MD, FACS, FAAP
Professor and Division Chief, Pediatric Otolaryngology
O T O LAR Y N G O LO G Y - H EAD & N E C K S U R G E R Y
University of Michigan Health System
CANC
deafness
diabetes
CV disease
Head/neck
breast
adrenal
colon
Types of deafness
• Conductive
• Sensorineural
• Auditory
neuropathy
• Syndromic vs.
nonsyndromic
Hearing Loss
• Most common sensory disorder, affects 28 million
Americans
• Over 500 syndromes such as Usher (deafness and
retinitis pigmentosa causing blindness)
• 50% develop presbycusis (age-related HL) by age 80
• 2 in 1000 children are born with educationally
significant hearing loss; 5 in 1000 by adolescence
• Association studies of presbycusis show increased
risk with smoking, high BMI, noise exposure;
moderate alcohol consumption is protective
• SNHL can be rehabilitated but not cured
The inner ear
•
•
•
•
Inaccessible, within the temporal bone
Tiny quantities of proteins
Genetics
Animal models
– Proteomics
– Histopathology
Richardson et al, Ann
Rev Physiol 2011
Positional cloning/ positional candidate
approach to discovering novel genes
Human
Genome
Working Draft
GENE1
GENE2
GENE3
……
GENE200
Dror and Avraham, Neuron 2010
DFNA2
DFNA37
DFNB32
DFNA7
DFNA49
DFNB36
DFNA34
DFNB45
DFNB7/11
DFNA47
DFNA51
DFNA36
DFNB31
DFNB33
DFNB3
DFNA20/26
2010
2012
DFNA43
DFNB9
DFNB47
DFNB58
DFNB27
DFNA16
DFNB59
DFNB30
DFNB23
DFNA19
DFNB12
DFNB57
DFNB19
DFNA6/14
DFNB6
DFNB42
DFNA18
DFNB15
DFNA44
DFNB51
DFNA32
DFNB18
DFNB63
DFNB2
DFNA11
DFNA8/12
DFNB21
DFNB24
DFNB20
DFNB68
DFNA57
DFNB15
DFNA4
DFNA38
DFNB25
DFNB55
DFNA27
DFNA24
DFNB26
DFNA39
DFNA42
DFNA52
DFNB49
DFNB60
DFNA1
DFNA15
DFNA42
DFNB66
DFNA13
DFNB53
DFNA22
DFNB37
DFNB31
DFNA21
DFNB38
DFNA10
DFNB44
DFNA5
DFNB39
DFNB4
DFNB14
DFNB17
DFNB13
DFNA28
DFNA50
DFNB62
DFNA31
DFNA48
DFNA25
DFNA41
DFNB50
DFNB1
DFNA3
AUNA1
DFNA9
DFNB5
DFNA23
DFNB35
DFNA53
DFNB16
DFNA30
DFNB48
DFNA40
DFNB22
DFN6
DFN4
DFNB65
DFNB8/10
DFNB29
DFNA17
DFNB28
DFNB40
DFN3
DFN2
DFNY
Courtesy Heidi Rehm, PhD
Connexin 26 (GJB2), DFNB1
• Mutations account for 15-50% of recessive
non-syndromic SNHL in US population
• Testing widely available ($300-$500)
• One mutation (35delG) accounts for 75-80%
of all mutations in Caucasians
• 35delG homozygotes typically have congenital
profound SNHL, but modifiers likely exist
• 3.01% carrier rate of 35delG in Iowa
population (Green et al. 1999)
35delG/+
+/+
Patterns of sensorineural hearing loss
Frequency in Hertz (Hz)
250
500
750 1000
1500 2000
0
X
O
Hearing Threshold (dB)
10
20
30
X
40
O
50
O
X
O
X
X
3000 4000
O
X
Frequency in Hertz (Hz)
6000 8000
-10
O
X
O
X
60
70
80
0
10
Hearing Threshold (dB)
-10
20
X
500
750 1000
1500 2000
3000 4000
][
O
X
O
X
O
30
O
X
40
[
][
X
O
50
O
X
6000 8000
O
X
O
X
O
X
60
70
90
80
100
90
Low frequency SNHL
WFS1 mutations
Also DIAPH1, FGFR3
250
O
[
X
]
Frequency (Hz) vs Right A/C
Frequency (Hz) vs Right B/C
Frequency (Hz) vs Left A/C
Frequency (Hz) vs Left B/C
100
Mid frequency “cookie-bite”
TECTA (alpha tectorin) mutations
High frequency or flat audiogram: many genes
Genetic testing for nonsyndromic
auditory neuropathy
• OTOF (otoferlin)
– 48 coding exons and splice sites; recessive
– $1500 through Partners- does not accept 3rd party
payment
• Pejvakin – reported in two families from Iran
• DIAPH3 (Schoen et al. 2010, PNAS)
– Autosomal dominant auditory neuropathy;
currently no lab offering clinical testing
BCN view of the world
• Does a definitive dx remain uncertain even after H&P
exam and conventional diagnostic studies?
• Does the patient display clinical features related to
the inherited mutation in question?
• What justifies the patient’s increased risk? i.e. family
history, ethnic background
• Is this disease treatable or preventable?
• Will the result of the test directly influence the tx
being delivered to the patient?
• Requires informed consent and genetic counseling
Deafness gene platforms
• OtoChip™- 19 assorted genes for deafness
including Usher, mitochondrial, $3800
• Otogenetics: 80+ genes, $488/sample
• OtoSCOPE: 60+ genes, $2000/sample
• Missing most “syndromic” genes such as:
– FGF3 : deafness, microtia, and microdontia (DFNB63)
– BSND: Bartter syndrome, DFNB73 (mild renal dysfunction)
– FGFR3: Muenke craniosynostosis; CATSHL (camptodactyly
and tall stature with HL)
Whole exome sequencing
• All coding exons and intron-exon junctions
• At least 2 samples/family to reduce number of
variants for follow-up
• Cross referencing to linkage region(s)
• Trending toward whole genome sequencing
(at some lower coverage)
– Unknown exons expressed specifically in the ear
– Splice mutations, distant enhancers/promoters
– Better sensitivity for deletions/duplications
Annotation
• dB SNP filter
– Includes deafness mutations such as 35delG
– Allow minor allele frequency <5%
• Exome variant server (U Washington) filter
• Up to 10,000 alleles
• Subjects not necessarily screened for deafness;
some reported deafness mutations seen
• Splice mutations may be called as non-synonomous
or synonomous variants
• Deafness mutations may occur in noncoding
sequence (5’UTR mutation in DIAPH3)
Interpretation
• Pooling samples reduces costs but may
sacrifice coverage of particular linkage region
• Failure to confirm on Sanger sequencing
• Easier to detect homozygous mutations
causing recessive disease
• Dominant requires analysis of every variant
and confirmation of segregation in the family
Most common non-genetic cause:
Human cytomegalovirus infection
•
•
•
•
HCMV affects 1.1/1000 live births
5% symptomatic; 30-65% incidence of SNHL
95% are asymptomatic; 8-15% incidence SNHL
Must detect in first 3 weeks of life; universal
screening of blood or saliva; newborn blood
spots have low sensitivity (JAMA 2010)
• Antibody testing may be able to exclude
HCMV as an etiology
The future is here!
• Identifying a genetic etiology
helps with prognosis and
recurrence risk
• Economies of scale will bring
the price down
• Developing a clinical algorithm
and drawing genotypephenotype correlations
Interval Genomic Viewer plots
3 SNPs in ZNF575
 Average 10x
 chr19:44039571
 chr19:44039572
 chr19:44039574
 Alignment error
 PCR duplication
artifact
 Note poor coverage
of 3rd sample

Jishu Xu
1 SNP in KCNN4





chr19:44271753
A19:4T (23x)
A34:6T (40x)
AA 14:0 (14x)
PCR duplication
artifact
Utility of chromosomal microarray analysis
(CMA) for evaluation of deafness
•
•
•
•
•
GJB2 null alleles due to distant deletions
~100kb deletion of STRC (DFNB16)
CHD7 (CHARGE syndrome)
TWIST (Saethre-Chotzen syndrome)
Recessive disorders with 0-1 mutations (e.g.,
SLC26A4 gene )
555kb deletion of chromosome 16p13.3
Non-syndromic Sensorineural Hearing Loss:
Summary
• Recessive is usually congenital and profound; GJB2
(also pseudodominant)
• Dominant is usually delayed onset and progressive
– Low frequency WFS1
– Mid frequency “cookie-bite” TECTA (alpha-tectorin)
• Auditory neuropathy: OTOF, DIAPH3
• Otherwise, no predominant genes; some in single
families or yet unidentified
• Next generation sequencing technologies
Genotype cohorts provide some prognostic
information
Snoeckx et al. GJB2 mutations and hearing loss: a
multicenter study, Am J Hum Genet, 2005.
OtoChipTM
• 19 genes, 70,000 bases; $3800, 8 week TAT
• Nonsyndromic SNHL: CDH23, DFNB31 (WHRN),
GJB6, MYO6, MYO7A, OTOF, PCDH15, SLC26A4 (PDS),
TMC1, TMIE, TMPRSS3, USH1C
• Mitochondrial tRNAser(UCN) and 6 in 12S rRNA
• OTOF, SLC26A4
• Usher syndrome : CDH23, CLRN1, DFNB31, GPR98
(exons 8, 20, 31-41 & 89), MYO7A, PCDH15, USH1C,
USH1G, USH2A
• Not included : WFS1, TECTA, KCNQ4
OtoSCOPE (Shearer et al. 2010)
• 50 nonsyndromic deafness genes, including
mitochondrial and miRNA, and 4 Usher genes
• Identified genetic etiology in 5/6 unknowns
– p.A366T/p. N1098S in CDH23 (DFNB12, USH1D)
– ~100kb deletion in STRC/p.Q1353X
– p.D288DfsX17 in MYO6 (DFNA22)
– p.E1965X in MYH14 (DFNA4)
– p.L281S in KCNQ4 (DFNA2)
– Found 4 variants in one dominant case, none
segregating with deafness- etiology unknown
AudioGene
http://audiogene.eng.uiowa.edu
• Input the audiometric thresholds and
AudioGene will predict the most likely genes
for dominant SNHL
KCNQ4
• Not available for all genes
• Can’t analyze more than
one audiogram at a time
• Can view audioprofiles for
selected genes
• Often, noise exposure
alters the classic pattern
High frequency SNHL
•
•
•
•
KCNQ4 mutation
Noise exposure
Presbycusis
Or many other genes!
1 SNP in SARS2

chr19:39408746