Characterization of Aryl Hydrocarbon Receptor
Interacting Protein (AIP) Mutations in Familial Isolated
Pituitary Adenoma Families
Susana Igreja,1yHarvinder S. Chahal,1yPeter King,1Graeme B. Bolger,2Umasuthan Srirangalingam,1Leonardo Guasti,1
J. Paul Chapple,1Giampaolo Trivellin,1Maria Gueorguiev,1Katie Guegan,3Karen Stals,3Bernard Khoo,4Ajith V. Kumar,5
Sian Ellard,3Ashley B. Grossman,1Ma ´rta Korbonits,1?and the International FIPA Consortiumz
1Department of Endocrinology, Barts and the London School of Medicine, Queen Mary University of London, London, United Kingdom;
2Comprehensive Cancer Center, University of Alabama, Birmingham, Alabama;3Department of Molecular Genetics, Royal Devon and Exeter
Foundation Trust, Exeter, United Kingdom;4Department of Endocrinology, UCL Medical School, Royal Free Campus, London, United Kingdom;
5North East Thames Regional Genetics Service, Great Ormond Street Hospital, London, WC1N 3JH, United Kingdom
Communicated by Georgia Chenevix-Trench
Received 9 February 2010; accepted revised manuscript 13 May 2010.
Published online 18 May 2010 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/humu.21292
ABSTRACT: Familial isolated pituitary adenoma (FIPA) is
an autosomal dominant condition with variable genetic
background and incomplete penetrance. Germline muta-
tions of the aryl hydrocarbon receptor interacting protein
(AIP) gene have been reported in 15–40% of FIPA
patients. Limited data are available on the functional
consequences of the mutations or regarding the regula-
tion of the AIP gene. We describe a large cohort of FIPA
families and characterize missense and silent mutations
using minigene constructs, luciferase and b-galactosidase
assays, as well as in silico predictions. Patients with AIP
mutations had a lower mean age at diagnosis (23.6711.2
years) than AIP mutation-negative patients (40.4714.5
years). A promoter mutation showed reduced in vitro
activity corresponding to lower mRNA expression in
patient samples. Stimulation of the protein kinase
A-pathway positively regulates the AIP promoter. Silent
mutations led to abnormal splicing resulting in truncated
protein or reduced AIP expression. A two-hybrid assay of
protein–protein interaction of all missense variants
showed variable disruption of AIP-phosphodiesterase-
4A5 binding. In summary, exonic, promoter, splice-site,
and large deletion mutations in AIP are implicated in 31%
of families in our FIPA cohort. Functional characteriza-
tion of AIP changes is important to identify the functional
impact of gene sequence variants.
Hum Mutat 31:950–960, 2010. & 2010 Wiley-Liss, Inc.
KEY WORDS: pituitary
AIP; tumor suppressor
The majority of pituitary adenomas are sporadic, but
occasionally they may also occur in a familial setting [Melmed,
2006]. Familial isolated pituitary adenoma (FIPA; MIM] 102200)
is an increasingly recognized autosomal dominant disease with
low or variable penetrance [Beckers and Daly, 2007; Chahal et al.,
2010]. Heterozygous germline mutations in the aryl hydrocarbon
receptor-interacting protein (AIP; MIM] 605555) have been
identified in 15–40% of FIPA families, whereas in the majority of
the families the disease-causing gene or genes are not known. AIP
mutation-positive families most commonly present with somato-
troph, or with both somatotroph and lactotroph adenomas, but
rarely other types of pituitary tumors can be observed [Daly et al.,
2007; Iwata et al., 2007; Leontiou et al., 2008; Toledo et al., 2007;
Vierimaa et al., 2006]. Young onset, seemingly sporadic somato-
troph adenoma patients can also harbor germline AIP mutations
[Georgitsi et al., 2008a]. Patients have loss of heterozygosity
(LOH) in the tumor tissue at the locus of the AIP gene at the
11q13 area [Gadelha et al., 2000; Soares et al., 2005; Vierimaa
et al., 2006; Yamada et al., 1997]. However, LOH at 11q13 has been
observed frequently in sporadic pituitary adenomas [Farrell and
Clayton, 2000], and there are a number of FIPA families with
11q13 LOH, but no detectable AIP mutations [Leontiou et al.,
2008; Soares et al., 2005]. Interestingly, the 11q13 locus also
contains the Multiple Endocrine Neoplasia type 1 gene (MEN1;
MIM] 131100), 2.4 megabases upstream from the AIP gene, which
can cause familial pituitary adenomas; however, the phenotype of
MEN1 and FIPA is different.
The AIP gene consists of six exons encoding a 330 amino acid
protein with three typical tetratricopeptide repeat (TPR) domains and
a final extended a-helix (a-7). AIP appears to function as a tumor
suppressor gene: we have previously reported data showing that
mutant AIP proteins lose the ability of wild-type (WT) AIP to
decrease cell proliferation and are unable to bind protein partners
[Leontiou et al., 2008]. Most of the described AIP mutations identified
via standard sequencing techniques change the amino acid sequence
due to nonsense, deletion, or insertion mutations, and result in a loss
of the C-terminal end of the protein. Large gene deletions also severely
disrupt the protein structure. Exon/intron junction mutations may
affect splicing or RNA stability and promoter mutations affect RNA
expression, but these have not been previously characterized.
& 2010 WILEY-LISS, INC.
Additional Supporting Information may be found in the online version of this article.
?Correspondence to: Ma ´rta Korbonits, Queen Mary University of London, Barts &
The London School of Medicine, Department of Endocrinology, Charterhouse Square,
London, UK, EC1M 6BQ. E-mail: email@example.com
yThese authors contributed equally to this work.
zSee list of consortium members in Appendix.
We have systematically examined our large cohort of patients
with FIPA specifically looking for large deletions, promoter, and
splice-site mutations in addition to exonic mutations; we describe
11 new AIP mutation-positive families with six novel mutations.
We present functional data on an AIP promoter mutation, a splice-
mutation and a synonymous change, which leads to reduced AIP
mRNA expression, as well as quantitative data on the effects of all
the published missense mutations on the interaction of AIP with
one of its binding proteins, phosphodiesterase-4A5 (PDE4A5).
These data allow us to characterize the clinical features of FIPA and
to provide the most accurate data for the functional impact of AIP
variants, which has an important role in genetic counseling of
patients and their families and will help inform the need for
predictive testing and biochemical and imaging screening.
Patients and Methods
We studied 38 novel families with FIPA (Supp. Table S1) who
were identified as having at least two family members with
pituitary adenoma and no features of MEN1 or Carney complex
(MIM] 160980). All patients provided written informed consent,
and institutional review board approval was obtained. For the
clinical analysis we combined these novel families with our
previously described cohort of 26 families (Supp. Table S2)
[Leontiou et al., 2008]. The clinical features of some of these
families have previously been described, as detailed in Supp. Table
S1 and S2 [Gadelha et al., 1999, 2000; Georgitsi et al., 2007, 2008b;
Leontiou et al., 2008; Matsuno et al., 1994; McCarthy et al., 1990;
Pestell et al., 1989; Soares et al., 2005]. Direct sequencing of AIP
(NM_003977.2) included the entire coding sequence, conserved
splice sites (from the conserved A of the upstream branch site to
110 downstream of each exon) and 1,200 base pairs of the promoter
region. Nucleotide numbering throughout the manuscript reflects
cDNA numbering with 11 corresponding to the A of the ATG
translation initiation codon in the reference sequence, according
to the guidelines of the Human Genome Variation Society
(www.hgvs.org/mutnomen). The initiation codon is codon 1.
AIP sequencing data were compared with Caucasian (n596) and
Japanese (n578) subjects from the general population, as
previously described [Leontiou et al., 2008]. Multiplex ligation-
dependent probe amplification (MLPA, P244-kit MRC-Holland,
Amsterdam, The Netherlands) dosage analysis was carried out to
look for partial or whole gene deletions in all the families that
tested negative by direct sequencing for germline AIP mutations
and for whom a suitable quality of DNA sample was available.
RNA Extraction, RT-PCR, Cell Culture
RNA from whole blood was extracted using PaxGene tube and
extraction kit (Qiagen, Crawley, UK). RNA from rat pituitary cell
line GH3 was extracted using the RNeasy Mini Kit (Qiagen). Both
protocols include a DNase step. RT-PCR on 1mg RNA was
performed as described previously [Leontiou et al., 2008]. Real-
time PCR was performed with the TaqMan system using ready-
made or custom-made (surrounding the c.807C4T mutation)
AIP probe-primer kits (ABI, Warrington, UK). Reactions were
performed in triplicate using b-actin or GAPDH as a house-
keeping gene. Data were analyzed using the standard curve
method. Rat somatomammotroph GH3 cell line was cultured and
transfected as previously described [Leontiou et al., 2008].
The c.[?270_?269CG4AA;?220G4A] changes in the promo-
ter area were identified in a Japanese somatotroph adenoma family
(Family VI Supp. Table S2). These changes were not detected in
Japanese (n578) and Caucasian (n596) individuals from the
general population or in any of the studied family members
(affected or unaffected AIP mutation carriers or their noncarrier
family members or AIP mutation negative patients, n5150) of our
familial cohort, or in any of our sporadic pituitary adenoma cases
studied (n598), altogether 844 chromosomes. The transcription
factor binding-site searching programs TESS (http://www.cbil.
regulation.com/pub/programs/alibaba2/index.html) were used to
identify possible bindingsite
c.[?270_?269CG4AA;?220G4A] sequence changes in the AIP
promoter area (Fig. 1A and B). A 1.2-kb (upstream from the start
codon) PCR fragment was amplified from gDNA (forward
50TACAACCTCCATCTCCTGGG30, reverse 50GAGTCCGGAAGTTG-
CCGAAA30). The WT fragment was cloned into the SacI and SpeIsites
of the promoterless firefly luciferase vector pGL3-basic (Promega,
Southampton, UK). Three mutated constructs were prepared from the
WT template using the QuikChange site-directed mutagenesis kit
(Stratagene, LaJolla, CA): a dibasic ?270_?269AA mutant, a single
?220A mutant and a construct containing both ?270_?269AA and
?220A mutations (Fig. 1C). To study basal and stimulated activity,
GH3 cells were transfected with the four different constructs and 48hr
later were treated with db-cAMP (2mM), forskolin (10mM), phorbol-
12-myristate 13-acetate (PMA, 20nM) and H89 (30mM) for 5hr.
A plasmid encoding renilla luciferase, pRL-CMV, was included to
control for transfection efficiency and promoter activity was
analyzed using the dual luciferase assay (Promega). Peripheral
blood-derived cDNA samples were studied from the two affected
patients and three ethnically matched controls.
Splicing Mutation c.249G4T
The ALAMUT program (http://www.interactive-biosoftware.-
com/alamut/doc/1.5/splicing.html) was used to predict whether
the novel synonymous sequence variant c.249G4T, p.5(affecting
the third nucleotide of codon 83 originally coding glycine)
changes the splicing characteristics of the AIP gene (Family 28,
Fig. 2A). The primers (forward 50GCGGATATCATCGCAAGACT30,
reverse 50CCTCATCTTCCACATGGAGA30) were designed on
exons 1 and 3 to detect aberrant splicing events.
Synonymous Mutation c.807C4T
The ALAMUT, the ESEfinder v.3.0 (http://rulai.cshl.edu/cgi-bin/
tools/ESE3/esefinder.cgi?process5home) and the RESCUE-ESE
(http://genes.mit.edu/burgelab/rescue-ese/) programs were used to
predict whether the synonymous c.807C4T, p.5(affecting the
third member of codon 269 originally coding phenylalanine)
sequence variant of exon 6 changes the splicing characteristics of
the AIP gene (Fig. 3A) [Leontiou et al., 2008]. Conventional and
real-time RT-PCRs were performed on patient blood-derived cDNA
using primers spanning the exon 5–6 junction to compare AIP
expression between patients carrying the mutation and control
individuals. To further characterize the splicing events around this
sequence abnormality we performed a minigene splicing study. To
construct the minigene, gDNA from a patient harboring the
c.807C4Tmutation was PCR amplified with primers covering part
of the exon 5, intron–exon region and part of exon 6 including the
HUMAN MUTATION, Vol. 31, No. 8, 950–960, 2010
mutation (forward 50AAGCTGGTGGTCGAGGAGTA30, reverse
50CAAAGTGCTGGAGCTGGAC30). WT and mutant minigene
constructs were prepared and fragments were cloned into the EcoRI
site of pcDNA3.11 (Invitrogen, Paisley, UK). GH3 cells were
transiently transfected for 24–48hr with WTand mutant minigenes
separately or together in equal amounts. Transfection efficiency was
followed by cotransfection with an expression plasmid coding for
green fluorescent protein (GFP, pEGFP-N1) at a pcDNA3.11with
insert:pEGFP-N1ratio of 1:10. Expression of the minigene tran-
scripts was detected using vector-specific primers.
Interactions between AIP and PDE4A5 were studied to clarify
the role of missense sequence changes in protein–protein binding.
PDE4A5 was cloned into the NotI-site of pLEXAN to generate a
LexA DNA-binding-domain fusion [Bolger et al., 2003; Leontiou
et al., 2008]. WT AIP and nine missense AIP sequence variants
(c.47G4A, p.R16H [Buchbinder et al., 2008; Cazabat et al., 2007;
Daly et al., 2007; Georgitsi et al., 2007; Yaneva et al., 2008];
c.145G4A, p.V49M [Iwata et al., 2007]; c.308A4G, p.K103R
[Beckers et al., 2008], c.713G4A, p.C238Y [Leontiou et al., 2008],
c.721A4G, p.K241E [Daly et al., 2007], c.769A4G, p.I257V
[Montanana et al., 2009], c.811C4T, p.R271W [Daly et al., 2007;
Igreja et al., 2010; Jennings et al., 2009], c.896C4T, p.A299V
[Georgitsi et al., 2007], the hotspot mutation c.911G4A, p.R304Q,
and a novel stop mutation (c.490C4T, p.Q164X) were cloned into
the NotI-site of pGADN to generate GAL4 activation-domain
fusions. Quantitative b-galactosidase assays were performed in the
Saccharomyces cerevisiae strain L40 by the method of Guarente
 using O-nitrophenyl-b-D-galactopyranoside as a substrate.
Each mutation was tested in at least two different yeast clones.
In Silico Analysis
Location of AIP missense mutations were compared to the
available consensus sequences of TPR proteins. A hypothetical
model of AIP was constructed using the Phyre (Protein
Homology/analogY Recognition Engine) program [Kelley and
Sternberg, 2009], based on the crystal structure of FKBP51
shown, half-filled symbols represents carrier subjects. B: Location of the two AIP promoter sequence changes. Numbers in boxes represent
exon numbers. Nucleotide numbering reflects cDNA numbering with 11 corresponding to the A of the ATG translation initiation codon in the
reference sequence. C: Schematic representation of the four AIP promoter constructs. D: The wild-type (WT) and the c.?220A single mutation
showed similar promoter activity measured by luciferase assay. The ?270_?269AA dibasic mutant and the double ?270_?269AA and ?220A
mutant constructs showed decreased promoter activity (???Po0.001 vs. WT). E: AIP promoter activity after treatment with db-cAMP, forskolin,
and PMA. The WT and the ?220A single mutant constructs showed increased promoter activity after treatment with db-cAMP and forskolin
compared to vehicle treatment (]Po0.05, ]]Po0.01). Treatments did not affect the promoter activity of the dibasic ?270_?269AA and double
?270_?269AA and ?220A mutant constructs. Following db-cAMP and forskolin there was a significantly lower promoter activity in the
?270_?269AA and ?270?269AA and ?220A mutant constructs compared to WT (???Po0.001,??Po0.01 vs. WT). F: The PKA inhibitor H89
inhibits the stimulating effect of forskolin on luciferase activity of the WT-promoter.
AIP promoter mutations. A: Family tree (Family VI, Supp. Table S2), filled circles represent patients with gigantism with age of onset
HUMAN MUTATION, Vol. 31, No. 8, 950–960, 2010
(MIM] 602623), the closest related protein with available crystal
structure (c1kt0A). The alignment is showing 22% identical amino
acids and 45% similar amino acids in the N-terminal region (amino
acids 3–90 of AIP) and the TPR domain area (amino acids 165–328
of AIP) (Supp. Fig. S1). Alignment was performed with CLUSTAL
2.0.12 (http://www.clustal.org/) and BL2SEQ (www.ncbi.nlm.
nih.gov/blast/bl2seq/wblast2.cgi) sequence alignment tools.
Statistical analysis was performed with StatsDirect software
(Addison-Wesley-Longman, Cambridge, UK). Distribution of
data was analyzed by the Shapiro-Wilk test. Comparisons were
calculated with the Student t-test and the Kruskal-Wallis test
followed by the Conover-Inman test, as appropriate. Data are
shown as mean7standard error, unless otherwise stated. Signi-
ficance was taken at Po0.05.
We report here 38 novel families with FIPA, including 11
families with an AIP mutation (Supp. Table S1). Six of these 11
families harbor a novel mutation including two complete gene
The novel mutations revealed via sequencing include a
nonsense mutation (c.490C4T, p.Q164X), a single base-pair
(c.249G4T, p.G83AfsX15, see data below) and a complex
deletion-insertion mutation (c.74_81delins7, p.L25PfsX130), all
resulting in a frameshift and/or premature stop-codon. Three
families have been found to have the previously well-described
c.910C4T, p.R304X hotspot [Daly et al., 2007; Jaffrain-Rea et al.,
2009; Leontiou et al., 2008; Vierimaa et al., 2006] mutation and
one family the c.911G4A, p.R304Q hotspot [Daly et al., 2007;
Leontiou et al., 2008] mutation. We also identified a previously
described variant (c.896C4T, p.A299V) in one of our families
that also harbors the p.R304X mutation.
We screened all affected patients in our cohort (Supp. Tables S1
and S2) for large deletions using MLPA, where no AIP mutations
were identified by sequencing and suitable DNA was available.
Three large deletions were identified among the 36 families (8.3%,
Family 9, Family IV, and Family XXI; Supp. Tables S1 and S2), all
three originally reported to have WT AIP sequences by conven-
tional sequencing [Georgitsi et al., 2007; Leontiou et al., 2008].
Two of these resulted in the loss of the full length AIP gene and
one was the previously described exon 2 deletion [Georgitsi et al.,
For clinical characteristics we analyzed the data from the
currently reported 38 families together with the 26 families we
have reported earlier [Leontiou et al., 2008]. In our cohort of 64
FIPA families with 160 affected subjects (Supp. Tables S1 and S2),
the mean age (7SD) at diagnosis of pituitary adenoma was
33.7715.5 years (23.6711.2 years in the AIP mutant families and
40.4714.5 years in the AIP negative families; Po0.00001). If we
recalculate the age of onset only for the successfully performed
MLPA negative families this was 38.4713.8 years (not signifi-
cantly different from the full cohort of AIP negative families).
Families with FIPA show variable penetrance. We are fully aware
that penetrance calculations are subject to bias due to sympto-
matic patient referral and due to incomplete genealogical data as
well as variable ages of the subjects studied, but these calculations
can still provide some useful information. Penetrance was
calculated in our families with taking into account affected and
obligate carrier subjects as well as half of the subjects with 50%
risk to inherit the mutation [Ogino et al., 2007]. We have a
mean7SD penetrance of 42721% for AIP-positive families with
a minimum of 8% and a maximum of 83%. The penetrance in the
AIP negative families was not significantly lower at 38716%,
minimum 19% and maximum 67%. However, we observed a
difference in the reported number of affected patients between
AIP mutation-positive and -negative families, suggesting a
difference in penetrance: the mean number of affected subjects
in AIP mutation-positive families was 3.271.8, whereas in AIP
mutation-negative families it was 2.270.4 (Po0.001). Interest-
ingly, 16 of our 20 families with an AIP mutation had at least one
member with gigantism and/or disease onset o18 years, whereas
only 3 of our 44 AIP-negative FIPA families had a member with
gigantism and/or disease onset o18 years. In the 20 AIP-positive
families 7 of the 63 affected subjects (11%) were diagnosed above
the age of 35 years, three out of these (5%) were above 50 years.
In the 44 AIP-negative families 56 of 96 patients (58%) were
diagnosed above the age of 35 years, 24 of these (25%) above the
age of 50 years. Five of 63 AIP mutation patients (8%) presented
with pituitary apoplexy; three of them were in childhood.
We studied whether the type of AIP mutation (presence or
the 249G4T mutation. A: Family tree (Family 28, Supp. Table S1)
showing patients with gigantism (filled square), prolactinoma (striped
squares), and carrier subjects (half-filled symbols) with age of onset
shown. B: Schematic representation of splicing in the WT and mutant
gene showing the location of the mutation (arrow) and the 32bp lost
from the end of exon 2 (shaded area) followed by a novel stop-codon
after 15 novel codons (marked with ?). Numbers in boxes represent
exon numbers. Primers used are shown by arrowheads. C: RT-PCR
using a patient and control cDNA. Patient cDNA shows an extra band
(425bp) that corresponds to the alternatively spliced AIP transcript
with the upper band showing the WT transcript (457bp). The identity
of these PCR products were confirmed by sequencing.
Alternatively spliced AIP transcript in the presence of
HUMAN MUTATION, Vol. 31, No. 8, 950–960, 2010
absence of C-terminal end of the protein) would influence
penetrance. We have 16 families with mutations resulting in no
protein or truncated proteins having 3.4371.9 patients per family
and four families with full-length proteins having 2.570.58
patients per family (no significant difference), suggesting the lack
of genotype–phenotype correlation in these 20 AIP-positive
families in terms of penetrance. Out of the 64 families, 59 had
at least one member with a prolactinoma or acromegaly, whereas
five families had pure nonfunctioning adenomas. Out of the 63
AIP mutation patients only 6 had ‘‘pure’’ prolactinoma, and all of
these were macroadenomas. The AIP mutation-positive patients
who clinically presented with a nonfunctioning adenoma all had
positive GH and/or PRL staining in the adenoma tissue,
suggesting that all subjects with AIP mutations have GH or
PRL-synthesizing adenomas in our cohort.
AIP Promoter Mutation Causes Decreased Promoter
Two mutations (c.[?270_?269CG4AA; ?220G4A] in cis)
had been identified in a Japanese FIPA family with two sisters
suffering from gigantism (Family VII; Supp. Table S2, Fig. 1A
and B) [Leontiou et al., 2008]. These changes are located 160 and
111bp upstream of the 50UTR. The transcription factor binding-
site searching programs suggested that the ?270_?269CG4AA
promoter mutation causes disruption of several transcription
factor binding sites. To investigate whether the observed sequence
changes affected the transcriptional activity of the AIP promoter,
we cloned a 1.2-kb upstream fragment of the AIP gene into the
pGL3-basic luciferase reporter vector (WT construct). To identify
which of the mutations were functional, promoter constructs were
prepared with a dibasic change (?270_?269AA), a single base
(?270_?269AA and ?220A) (Fig. 1C). These were transiently
transfected into GH3 cells and the activity of the promoter was
measured using a luciferase assay.
Transfection of the WT promoter and the single ?220A promoter
construct resulted in an 18-fold increase of promoter activity
compared to empty vector pGL3-basic (Fig. 1D). Compared to the
WT promoter construct the dibasic ?270_?269AA promoter
construct and the double ?270_?269AA and ?220A promoter
construct had significantly decreased promoter activities (Po0.001).
asymptomatic carrier with age of onset of disease (Family XVI, Supp. Table S2). The location of the mutation is shown by an arrow; numbers in
boxes represent exon numbers. B: Conventional RT-PCR with AIP primers on blood-derived cDNA obtained from three control individuals (C1–3)
and two subjects (M1 and 2) carrying the 807C4TAIP mutation. Patients carrying the 807C4Tshowed decreased AIP expression. C: Real-time
PCR using an AIP primer and probe set to compare the AIP levels between control individuals and patients; C (control). D: Wild-type (WT) and
mutant minigene constructs. E: Conventional RT-PCR using vector-specific primers for the minigene constructs showing decreased mutated
minigene expression (807T) and intermediate level of expression for the coexpression of WT and mutant minigenes (807CT). F: Real-time PCR
showing an increased AIP expression in the WT minigene construct compared to mutant and coexpression of WT and mutant minigenes. EV,
Decreased AIP mRNA expression in the presence of the 807C4T. A: Family tree showing two patients with acromegaly and an
HUMAN MUTATION, Vol. 31, No. 8, 950–960, 2010
These results suggest that the dibasic ?270_?269AA promoter
mutation, but not the single ?220A variant, reduces promoter
activity, which is likely to have a functional impact in the patients
carrying this dibasic mutation, and suggests that this area is
important for the regulation of the AIP promoter.
Additional support for the reduced function of this AIP
promoter was generated with real-time PCR study of the two
affected subjects’ leukocyte cDNA (heterozygote for the promoter
mutation) compared to three healthy controls from the same
ethnic background. This showed reduced AIP expression (AIP/
GAPDH mRNA ratio, mean7SEM, 0.8670.2 vs. 0.5870.1,
P50.05), supporting the reduced activity of the AIP promoter in
AIP Promoter is Regulated via the cAMP–PKA Signaling
The cyclic AMP pathway is important for somatotroph
pathogenesis as its upregulation via somatic gsp mutations (MIM]
139320) or via PRKAR1A (MIM# 188830) mutations lead to
somatotropinomas [Horvath and Stratakis, 2008]. cAMP stimu-
lates AIP binding partner AhR [Oesch-Bartlomowicz et al., 2005].
In addition, the extracellular signal-regulated kinase (ERK)
pathway has recently been shown to be deregulated in pituitary
adenomas [Dworakowska et al., 2009]. No previous data are
available about the regulation of the AIP promoter. We therefore
treated GH3 cells transfected with the 1.2-kb AIP promoter
constructs with db-cAMP (a cAMP analog which activates PKA),
forskolin (an adenylate cyclase activator that activates PKA), and
PMA (a protein kinase C activator leading to activation of the
ERK/MAPK pathway) and measured luciferase activity. Treatment
with db-cAMP and forskolin increased WT and ?220A promoter
activity compared to vehicle (Fig. 1E). The effect of forskolin on
the WT promoter was inhibited by the PKA inhibitor H89
(Fig. 1F). However, db-cAMP and forskolin treatment did not
have any effect on the dibasic ?270_?269AA mutated promoter
or the double ?270_?269AA and ?220A mutated promoter.
Treatment with PMA did not alter AIP promoter activity in any of
the four promoter constructs. These results implicate the
involvement of the cAMP–PKA signaling pathway in regulating
AIP expression, and suggest that the ?270_?269 region is
required for cAMP-induced AIP promoter activity.
Splice Mutation c.249G4T Causes Truncated AIP
A novel mutation (c.249G4T) was identified in a family
(Family 28; Supp. Table S1) with two childhood-onset pituitary
adenomas (a somatotrophinoma and a prolactinoma) and an
adult-onset prolactinoma (Fig. 2A). The mutation at the end of
exon 2 did not change the amino acid sequence, but in silico
analysis using ALAMUT predicted that this mutation created a
novel 50splice-site 32 bases upstream from the normal splice site
resulting in a frameshift and a stop-codon after 35 novel amino
acids (Fig. 2B). To determine whether this mutation indeed causes
alternative splicing, we extracted RNA from blood of one of the
patients. RT-PCR with primers upstream and downstream of the
mutation was performed and compared to cDNA from a control
individual. RT-PCR of cDNA from the patient heterozygous for
the mutation generated the expected WT product plus an
additional band that corresponded to the size of the predicted
alternative product (425bp; Fig. 2C). Sequencing of this smaller
product confirmed the identity of the alternatively spliced
transcript. These results confirmed the in silico prediction that
the nucleotide change c.249G4T causes alternative splicing
leading toa frameshift and
a truncatedAIP protein
The Synonymous c.807C4T Mutation Results in Reduced
The mutation c.807C4T, p.5at the beginning of exon 6 had
been identified in a family with two somatotrophinomas (Family
XVI; Supp. Table S2) [Leontiou et al., 2008]. In silico analysis
(ESEfinder and RESCUE-ESE) suggested that the c.807C4T
mutation might result in the loss of a binding site for splice
enhancers SRp40 and SRp55 (splicing regulatory proteins that
recognize exonic splicing enhancer sequences) in the final exon of AIP
(Fig. 3A). We explored whether this mutation affected the splicing
of the last AIP exon. cDNA obtained from affected or carrier
individuals was amplified with different primer sets covering the
various sections of the exon 5–30UTR region and compared to the
WT gene. PCR amplification with the different primer sets did not
show any novel alternatively spliced transcripts, but instead,
patients with the mutation showed decreased AIP mRNA
expression compared to control individuals using both conven-
tional (Fig. 3B) and real-time PCR (Fig. 3C). Thus, these data
suggests that the c.807C4T mutation may have a functional
impact by reducing AIP transcript levels.
To further investigate whether the c.807C4Tmutation leads to
decreased mRNA expression, we performed in vitro studies using
a minigene construct. The minigene was constructed with an
amplified DNA fragment from a patient carrying the mutation.
AWT (807C) and a mutated (807T) minigene including the 30end
of exon 5, the intron between exon 5 and 6, and the 50-end of exon
6 were cloned into the pcDNA3.1(1) expression vector (Fig. 3D).
The WTminigene and the mutated minigene were then transiently
transfected into GH3 cells for 48hr. To mimic the patient’s
heterozygous mutation, cells were cotransfected with equal
amounts of the WT and mutated minigenes. Using conventional
(Fig. 3E) and real-time PCR (Fig. 3F), we detected decreased
transcriptional products of the minigene in the cells transfected
with the mutated minigene (807T) and an intermediate amount
with the cotransfection of WT and mutated minigenes compared
to cells with the WT minigene construct. These data corroborate
the in vivo data showing that the c.807C4T variant is associated
with reduced levels of AIP mRNA, supporting the functional
impact of the mutation.
Does the c.896C4T, p.A299V Variant have a Functional
In Family 10, two AIP changes were identified in two
asymptomatic male carriers (70 and 65 years): the c.910C4T,
p.R304X (a known pathogenic mutation) [Leontiou et al., 2008]
and the c.896C4T, p.A299V (previously reported in a patient
with sporadic acromegaly [Georgitsi et al., 2007], but in none of
the general population subjects studied by us or others) variants.
Other family members showed only the c.910C4T or the
c.896C4T change. We cloned this region of the AIP gene from
the gDNA of a subject carrying both mutations and sequenced
several colonies. Some of the colonies carried the c.910C4T
change while others the c.896C4T change. These data confirm
that the two sequence changes reside on different chromosomes
(in trans), and therefore the asymptomatic subjects are compound
heterozygotes for these two AIP changes. Interestingly, one of the
female carriers of the p.A299V variant, but not the p.R304X
HUMAN MUTATION, Vol. 31, No. 8, 950–960, 2010
mutation, was diagnosed with a microprolactinoma at the age 30
years. Out of our seven ‘‘pure’’ prolactinoma patients found in our
AIP mutation-positive families she is the only one with a
microprolactinoma, and most of the others needed surgery and
radiotherapy. We consider that her case might be a phenocopy,
similar to the phenocopies reported in the Finnish Q14X
mutation family [Vierimaa et al., 2006], however, as she carries
the p.A299V AIP variant this is uncertain. We performed a
functional assay to study the interaction of AIP and PDE4A5. The
p.A299V variant of AIP did not show a profoundly reduced
binding to PDE4A5 compared to WT controls (Fig. 4A). In
summary, the clinical data on the compound heterozygote
subjects support that the c.896C4T, p.A299V change might be
a rare polymorphism and the in vitro data are compatible with this.
Analysis of Missense AIP Mutations
We decided to study the protein–protein interaction properties
of all missense mutations reported in the literature. We
have previously shown with a filter b-galactosidase assay that
the R81X, R217X, and R304X nonsense mutations disrupt
the binding between AIP and PDE4A5. Here we used a
of AIP with PDE4A5. It shows more than fivefold difference (activity 0–20% of wild-type [WT]) from WTAIP for missense mutations K103R, C238Y,
K241E, and R271W as well as for the positive control truncation mutation Q164X. The R16H, V49M, I257V, A299V, and the R304Q variants show no
difference or activity 33–100% of wild type (less than threefold difference) from WT AIP (mean7SD). b-Galactosidase activity was measured
as described by Guarente (1983). Each mutation was tested in at least two different yeast clones, with identical results (n53 for each clone).
B: Hypothetical structure of AIP based on the structure of FKBP51 showing the three tetratricopeptide (TPR) domains with three pair of
antiparallel a-helices and the final extended a-helix, a-7 (courtesy of Prof. David Barford, London, UK). Sequence comparison of human AIP and
FKBP51 is shown in Supp. Figure S1. Amino acids with reported missense variants are highlighted. C: The three TPR domains, each consisting of
an A and B helix, are shown of several Hsp90 binding proteins including AIP (table modified from Hidalgo-de-Quintana et al. ). Numbers in
diamond shapes show TPR motif position numbers. Amino acids marked with blue bold letters are important for the packaging and stability of
the a-helices. Amino acids at position 8 and 20 were shown to be important in helix A and B packaging, while position 27 helps packaging of
helix B with helix A of the same TPR domain and helix A of the following TPR domain. Underlined amino acids are predicted to be involved in the
peptide binding pocket of FKBP51. Full-length variants of AIP affecting the TPR domains are shown on the figure (see review [Tahir et al., 2010]).
Amino acids circled with red in the AIP sequence show missense variants described (C238Y, K241E, I257V, R271W). Amino acids marked with
orange bold italics have been shown to be replaced by stop codon in FIPA patients (K201X, Q239X, K241X, Y268X). The amino acid marked by a
red arrow is followed by an in-frame insertion (p.F269_H275dup) in a FIPA family. In-frame deletion variants p.Y238del is shown by strikethrough
and complex missense and in-frame deletion mutation p.[E293G; L294_A297del] is shown by black circle and strikethrough.
Single amino acid substitutions in AIP. A: A yeast two-hybrid quantitative b-galactosidase assay was used to assess the interaction
HUMAN MUTATION, Vol. 31, No. 8, 950–960, 2010
quantitative b-galactosidase assay to study all nine published
missense mutations. Figure 4A shows that the K103R, C238Y,
K241E, and R271W mutations completely disrupt the binding
with values fivefold less than WT, similar to that of the novel
Q164X nonsense mutation (Po0.01 for each comparison). The
V49M change showed no effect (mean7SD, 9771% of WT),
whereas the R16H, I257V, A299V, and R304Q variants showed a
small drop in binding compared to WT (3978%, 35725%,
58713%, and 76717% of WT, respectively).
We used a combination of clinical, experimental, and in silico
methods to analyze how AIP mutations would change the AIP
protein structure and function. Out of the known 49 AIP variants
35 result in a truncated or no protein (12 nonsense mutations, 12
frameshifts, six splice site, four large deletions, and one promoter
mutation) [Tahir et al., 2010]. Previous studies convincingly
showed that loss of the C-terminal end of the protein results in loss
of function [Leontiou et al., 2008; Petrulis and Perdew, 2002]. The
in-frame insertion mutation of exon 6 has been shown to disrupt
the antiproliferative effect of AIP [Leontiou et al., 2008]. No
functional data are available for the four in-frame deletions. The
nine missense changes result in full-length proteins (R16H, V49M,
K103R, C238Y, K241E, I257V, R271W, A299V, and R304Q). The
location of these missense mutations are shown on a model of AIP,
which was based on the structure of FKBP51 (Fig. 4B).
R16H has been identified in control subjects as well as sporadic
and familial patients, and no LOH was shown in a tumor sample
[Buchbinder et al., 2008; Cazabat et al., 2007; Daly et al., 2007;
Georgitsi et al., 2007; Yaneva et al., 2008]. A reduction in PDE4A5
binding was observed, but this was less than threefold difference
from WT. It is important to note that the PDE4A5-AIP interaction
is at the C-terminal end of the molecule [Bolger et al., 2003].
Although the R16 amino acid is conserved in vertebrates, the
R16H change probably represents a rare polymorphism as
previously suggested [Raitila et al., 2007].
The V49M change was found in a sporadic childhood-onset
somatotroph adenoma case, but no LOH was found in the tumor
sample [Iwata et al., 2007]. The PDE4A5 binding data was similar
to WT, as expected, as PDE4A5 binds the C-terminal of the AIP
molecule while this variant affects the N-terminal of the protein.
This change is conserved in most vertebrates, but not in lower
species. Although no further data are available, the clinical
significance of this variant is uncertain.
The K103R variant was identified in a patient with a childhood-
onset corticotroph adenoma [Beckers et al., 2008]. Although the
K4R is conservative substitution, the K103 residue is a conserved
amino acid in most species and the PDE binding assay suggests a
profoundly reduced activity (875% of WT), supporting the
possibility that this mutation has a functional impact.
The C238Y mutation was found in a Mexican family with three
acromegalic brothers [Leontiou et al., 2008]. It disrupts a
consensus TPR motif amino acid. To understand the relevance
of this amino acid change we need to examine the structure of the
TPR domains. AIP has three TPR motifs (Fig. 4B and C). Each
TPR motif is composed of a pair of antiparallel helices, termed
helices A and B. The consecutive helices of TPR domains are
packaged together so each helix shares two immediate helix
neighbors [Das et al., 1998]. A typical TPR consensus sequence
pattern of small and large hydrophobic residues has been defined
[D’Andrea and Regan, 2003; Sikorski et al., 1990]. Consensus
amino acids (TPR residue positions marked as numbered
diamonds on Fig. 4C) are located at position 4, 7, 8, and 11 in
helix A and position 20, 24, 27, and 32 in helix B [D’Andrea and
Regan, 2003; Sikorski et al., 1990]. Small hydrophobic residues are
commonly observed at positions 8, 20, and 27 within the TPR
motif. Residues 8 and 20 are located at the position of closest
contact between the A and B helices of a TPR, whereas residue 27
on helix B is located at the interface of three helices (A, B, and the
A helix of the next TPR motif) within a three-helix bundle [Das
et al., 1998]. The C238Y mutation is at position 8 of the A helix of
the second TPR domain of AIP and its mutation could destabilize
the A and B helix packaging of the second TPR domain (Fig. 4C)
[D’Andrea and Regan, 2003; Hidalgo-de-Quintana et al., 2008].
Both the cell proliferation [Leontiou et al., 2008] and the PDE
binding assays (Fig. 4A) show that this mutant loses its activity in
these two different types of functional assays. Therefore, all the
data strongly suggest that this mutation has a functional impact.
The K241E mutation was identified in a FIPA family [Daly
et al., 2007]. Interestingly, the same amino acid can also be
affected by a stop mutation [Beckers et al., 2008]. This large
conserved residue is part of the TPR sequence consensus at
position 11 (Fig. 4C), and is suggested to be required for the
packing of adjacent TPR a-helices [D’Andrea and Regan, 2003;
Hidalgo-de-Quintana et al., 2008]. Our experimental data on
protein interaction show a reduced ability to interact with
PDE4A5 supporting the possible functional impact of the variant.
The I257V change affects a conserved small residue in most
vertebrates in the B helix of the second TPR domain at position
27, which is thought to be important in the packaging of the
adjacent helices [Das et al., 1998], in this case, A and B helices of
TPR2 and A helix of TPR3 (Fig. 4C). It was found in a sporadic
patient with unusual cell type (TSHoma) [Montanana et al.,
2009]. A reduction in PDE4A5 binding was observed, but this was
less than threefold difference from WT. LOH data are not
available; therefore, the possibility that it has a functional impact
cannot be excluded.
The R271W is the second most common mutational hotspot in
the AIP gene. This mutation has been identified in three
independent families [Daly et al., 2007; Jennings et al., 2009] as
well as in a sporadic giant [Igreja et al., 2010]. It disrupts a
conserved amino acid in the third TPR; it corresponds to an
arginine in several other TPR proteins at the equivalent location
such as in protein phosphatase 5 (PP5A), AIP-L1, FKBP51, and
FKBP52 proteins (Fig. 4C). Interestingly, this amino acid was
selected for mutation to alanine experimentally even before
human mutations were identified in this gene [Bolger et al., 2003;
Laenger et al., 2009]. Both the human mutation R4W (Fig. 4A)
and the experimental mutation of R4A [Bolger et al., 2003]
completely disrupts the AIP–PDE binding. These experimental
and clinical data strongly support the pathogenic role of this
The A299V change was described in a sporadic acromegaly
patient [Georgitsi et al., 2007] and in the current manuscript (see
above). This variant, which is conserved in most vertebrates, is
located at the beginning of the a-7 helix in the area that is relevant
for PDE4A5 interaction but binding in vitro was less than
threefold difference from WT. We identified this change in four
subjects of a family where the R304X mutation was also detected.
Two unaffected patients carried both changes; one unaffected
patient carried only the A299V change while one young female
carried only the A299V change and was diagnosed with a
microprolactinoma at the age of 30 years. As AIP knockout mice
are not viable [Lin et al., 2007], and the subjects with the
compound heterozygote genotype with the certainly pathogenic
R304X mutation are unaffected, this variant is unlikely to have a
functional impact, although the patient with a possible phenocopy
(small prolactinoma in a young female is not typical of AIP
HUMAN MUTATION, Vol. 31, No. 8, 950–960, 2010
mutation patients) carrying this change makes the status of this
The R304Q locus is part of a CpG island mutational hotspot.
This amino acid change is relatively conservative changing a
longer side chain positively charged amino acid to a slightly
shorter, uncharged, hydrophilic residue in the a-7 helix. The
amino acid is conserved in mammals, but not in lower species.
Our protein interaction data show a less than threefold difference
in PDE4A5 binding. The pathogenic role of R304Q, however, is
overwhelmingly supported by the fact that it has been identified in
several independent FIPA families as well as in sporadic patients
[Cazabat et al., 2007; Georgitsi et al., 2007; Leontiou et al., 2008;
Vargiolu et al., 2009] (and the current study).
In summary, our two-hybrid analysis of the missense mutations
showedthat they fellinto
were generally considered to have a functional impact had
b-galactosidase activity values more than fivefold different from
WT, whereas the mutants not generally considered to have a
functional impact had b-galactosidase values less than threefold
different from WT. Of note, none of the b-galactosidase values for
the missense mutants fell between these two values, providing a
clear distinction between functionally significant and functionally
nonsignificant mutants in this PDE4A5-related assay.
two groups:mutations that
We have analyzed clinical data and performed genetic and
functional studies to investigate the possible functional impact of
some of the reported AIP sequence changes. The main findings of
this study are: (1) all our AIP-positive families have at least one
member with GH/PRL-staining adenomas, have significantly
higher number of subjects affected suggesting higher penetrance
than in AIP-negative FIPA families, and have at least one
childhood-onset case in the majority of the families. (2) Six novel
AIP germline and five previously described mutations were
identified using sequencing and MLPA, and our data support
the suggestion that in addition to exon and exon–intron junction
sequencing, study of the promoter area, and a technique suitable
for the detection of large deletions, should be part of the DNA
analysis of patients with FIPA. (3) Our data suggest a functional
impact for the dibasic promoter mutation we have identified, as it
causes decreased promoter activity compared to the WT
promoter. In addition, we reveal that the cAMP–PKA signaling
pathway positively regulates the activity of the AIP promoter.
(4) The exonic synonymous mutations studied were shown to have
a functional impact due to abnormal splicing or decreased AIP
mRNA expression. (5) We combined clinical, in silico prediction
and protein–protein binding assay data to characterize all the
missense AIP changes identified in pituitary adenoma patients to
help define those that may represent rare single nucleotide
polymorphisms, and which may have a role in the disease process.
Previously described mutations affecting exon–intron junctions
were hypothesized to cause abnormal AIP protein based on in
silico predictions. Here we show actual data from patient cDNA
samples regarding the consequences of the c.249G4T mutation,
confirming the possible functional impact of this change. Studying
another silent mutation (c.807C4T), we predicted the abnormal
splicing of exon 6 due to the disrupted binding of the splicing
regulatory proteins. As this abnormality affects the last exon,
studying its role is not as straight forward as other splicing
mutations at earlier exons. Although our search for alternative
splice products using various sets of oligonucleotides did not
reveal an alternative splice product, a minigene approach
confirmed the lack of a normal exon 5–6 product and patient
cDNA samples showed reduced AIP mRNA expression. We
hypothesize that this mutation may either cause destabilization
of the mRNA leading to reduced expression levels, or alternative
splicing to a new isoform not detected by our RT-PCR assays.
AIP promoter regulation has not previously been studied. We
suggest that the c.?270_?269 area is part of the basic promoter,
as its disruption significantly reduces basal promoter activity in
vitro. Based on prediction programs, the mutation at c.?270_?269
position would disrupt several transcription factor binding sites,
including the zinc-finger protein SP1. SP1 can be activated by the
cAMP–PKA pathway [Rohlff et al., 1997], which is important in
proliferation and hormone release of somatotroph cells: gain-of-
function mutations in the a-subunit of the G-protein (GNAS) and
loss-of-function mutations of the regulatory subunit of PKA
(PRKAR1A) lead to deregulation of the cAMP–PKA pathway and
play a role in somatotroph tumorigenesis. Here we show that the
activity of the AIP promoter can be upregulated by the stimulation
of this pathway and the effect is inhibited by a PKA inhibitor. This
first appears to be a paradox, as sporadic somatotroph adenomas
show an overactive PKA pathway [Bertherat et al., 1995]. However,
our data are compatible with increased AIP mRNA and protein
expression described previously [Jaffrain-Rea et al., 2009;
Leontiou et al., 2008] and corresponds to the fact that low AIP
levels do not play a role in sporadic somatotroph tumorigenesis
when AIP mutations are not present [Leontiou et al., 2008].
PDEs can also regulate the PKA pathway, and PDE4A5 and
PDE2A are known partners of AIP [Bolger et al., 2003; de Oliveira
et al., 2007]. The AIP–PDE4A5 interaction was shown to be lost in
the presence of AIP mutations [Leontiou et al., 2008]. Here we
further studied all the described missense AIP changes as well as
our novel nonsense mutation with a quantitative assay. We found
that a stop mutation and four missense mutations (K103R, C238Y,
K241E, and R271W) completely disrupted the PDE4A5 binding.
Three of these affected amino acids are known to be important for
TPR structure [Das et al., 1998], and two have been experimen-
tally shown to disrupt the function of the AIP protein [Bolger
et al., 2003; Laenger et al., 2009; Leontiou et al., 2008]. Clinical
data suggest, although do not prove, that the R16H (found in
normals as well), V49M (lack of LOH in tumor sample), and the
A229V (also identified in unaffected subjects with compound
heterozygosity for the pathogenic R304X mutation) changes
might not have a functional impact, and this is supported by the
weak or no effect on PDE4A5 binding. The I257V change affects a
conserved amino acid at a location known to be important for the
a-helix packaging of the TPR domains, and despite the weak effect
on PDE4A5 binding a functional impact is possible. The
pathogenic role of the R304Q variant is beyond doubt. The lack
of profound reduction in PDE4A5 binding suggests that this part
of the molecule may not be important in this PDE4A5–AIP
interaction. On the other hand, missense variants can often act by
altering the stability or folding the protein, rather than affecting
amino acids directly involved in a protein–protein interaction.
Clearly, the lack of profoundly reduced binding to PDE4A5 does
not rule out other types of dysfunction of these variant AIP
molecules, and further studies are needed to clarify whether the
reduced or lack of AIP–PDE4A5 interaction as measured in this in
vitro assay is important for the pituitary pathogenic process.
We identified two asymptomatic carrier subjects with two
changes in the AIP gene—p.R304X and p.A299V—and showed
that these variants are located on opposite parental chromosomes.
It has previously been demonstrated that the p.R304X hotspot
mutation disrupts the function of the AIP molecule. The p.A299V
HUMAN MUTATION, Vol. 31, No. 8, 950–960, 2010
change (first described in a sporadic patient with acromegaly)
[Georgitsi et al., 2007] involves an amino acid substitution of
alanine to valine. This residue is not part of the TPR-domain
consensus. The PDE4A5 binding data did not support a profound
loss of function due to this missense mutation. Also, if this change
has functional impact, our subjects (65 and 70 years, both
asymptomatic) would be compound-heterozygotes for AIP
mutations, whereas AIP knockout mice are not viable (they die
at embryonic age e10.5–e14.5 due to cardiac malformations) [Lin
et al., 2007], thus suggesting this change has no functional impact.
We, however, identified a p.A299V female carrier with a
microprolactinoma. Combining all these data with the fact that
her phenotype is not typical of AIP mutation positive patients, her
case may represent a phenocopy. The A299V change is a variant
with unknow significance, and further studies are needed to
confirm that it has no functional impact.
More than two-thirds of FIPA families do not have a detectable
mutation in the AIP gene, and show differences in age of onset
and level of penetrance as well as heterogeneity of tumor types
compared to AIP positive patients, suggesting that there is a
phenotypic difference between FIPA families with AIP mutations
and FIPA families where the disease is caused by another gene or
genes. Currently, work is being carried out to identify these new
In summary, 31% of our FIPA families harbor AIP mutations.
They show a predilection for GH/PRL-synthesizing adenomas,
younger age of onset, and higher penetrance of disease compared
to AIP-negative FIPA families. The AIP promoter is positively
regulated by the cAMP–PKA pathway, and functional character-
ization of AIP mutations reveal that our promoter mutation
decreases promoter activity, whereas some synonymous changes
result in abnormal splicing or reduced AIP expression. Searching
for large deletions using MLPA and analysis of clinical data and
functional characterization of rare AIP changes helps to clarify the
pathogenic role of AIP in patients with familial and sporadic
pituitary adenomas, and assists in the appropriate genetic
counseling and clinical management of these patients and their
We are very grateful for the patients and family members who provided
samples. We thank Profs. David Barford and Carol Shoulders, as well as
Drs. Chrysanthia Leontiou and Lou Metherell, for the helpful comments.
HSC was supported by an MRC Training Fellowship. Projects in Prof
Korbonits laboratory are supported by the Cancer Research Committee of
St. Bartholomew’s Hospital, by the Wellcome Trust and by the NIHR.
Studies in Dr. Bolger’s laboratory are supported by the Bolger Prostate
Cancer Research Fund.
Appendix: List of International FIPA Consortium
Akker, Scott, London, UK; Atkinson, Brew, Belfast, Northern
Ireland; Aylwin, Simon, London, UK; Baldeweg, Stephanie,
London, UK; Bevan, John, Aberdeen, UK; Cheetham, Tim,
Newcastle, UK; Chew, Shern, London, UK; Choudry, Kiran,
Texas, USA; Clayton, Richard, Stoke-On-Trent, UK; Damjanovic,
Svetozar, Belgrade, Serbia; Darzy, Ken, Milton Keynes, UK;
Dattani, Mehul, London, UK; Davis, Julian, Manchester, UK;
Drake, Will, London, UK; Dzeranova, Larisa, Moscow, Russia;
Ede ´n, Britt, Engstro ¨m, Sweden; Eguchi, Kuniki, Hiroshima, Japan;
Fica, Simona, Bucharest, Romania; Flanagan, Daniel, Plymouth,
UK; Frohman, Lawrence, Chicago, Illinois, USA; Gadelha,
Monica, Rio de Janeiro, Brazil; Gallego, Patricia, Brisbaine,
Australia; Gla ´z, Edit, Budapest, Hungary; Goldman, James,
Boston, Massachustts, USA; Goldstone, Tony, London, UK;
Howlett, Trevor, Leicester, UK; Inder, Warrick, Melbourne,
Australia; Iwata, Takeo, Tokushima, Japan; Kaplan, Felicity,
Stevenage, UK; Karavitaki, Niki, Oxford, UK; Laws, Ed, Boston,
Massachusetts, USA; Lechan, Ron, Boston, Massachusetts, USA;
Levy, Miles, Leicester, UK; Matsuno, Akira, Chiba, Japan; Miljic,
Dragana, Belgrade, Serbia; Modenesi, Silvia, Belo Horizonte,
Brazil; Molitch, Mark, Chicago, Illinois, USA; Musat, Ma ˆda ˆlina,
Bucharest, Romania; Orme, Steve, Leeds, UK; Pato ´cs, Attila,
Budapest, Hungary; Popovic, Vera, Belgrade, Serbia; Powell,
Michael, London, UK; Quinton, Richard, Newcastle, UK;
Randeva, Harpal, Warwick, UK; Ribeiro de Oliveira JR, Anto ˆnio,
Belo Horizonte, Brazil; Scho ¨fl, Christof, Erlangen, Germany;
Soares, Beatriz, Belo Horizonte, Brazil; Spada, Anna, Milan, Italy;
Strasburger, Christian, Berlin, Germany; Swords, Francesca,
Norwich, UK; Tsagarakis, Stylianos, Athens, Greece; Vaks,
Vladimir, Oxford, UK; Wass, John A.H., Oxford, UK; Widell,
Hakan, Gotheborg, Sweden; Yarman, Sema, Istanbul, Turkey;
Yoshimoto, Katsuhiko, Tokushima, Japan.
Beckers A, Daly AF. 2007. The clinical, pathological, and genetic features of familial
isolated pituitary adenomas. Eur J Endocrinol 157:371–382.
Beckers A, Vanbellinghen JF, Boikos S, Martari M, Verma S, Daly AF, Raygada M,
Keil M, Papademetriou J, Drori-Herishanu L, Horvath A, Nesterova M,
Tichomirowa MA, Bours V, Marx S, Agarwal SK, Salvatori R, Stratakis CA. 2008.
Germline AIP, MEN1, PRKAR1A, CDKN1B (p27Kip1) and CDKN2C (p18INK4c)
gene mutations in a large cohort of pediatric patients with pituitary adenomas
occurring in isolation or with associated syndromic features. Proc of the 90th
Annual Meet of the Endocrine Soc; OR38-1.
Bertherat J, Chanson P, Montminy M. 1995. The cyclic adenosine 30,50-monophosphate-
responsive factor CREB is constitutively activated in human somatotroph
adenomas. Mol Endocrinol 9:777–783.
Bolger GB, Peden AH, Steele MR, MacKenzie C, McEwan DG, Wallace DA, Huston E,
Baillie GS, Houslay MD. 2003. Attenuation of the activity of the cAMP-specific
phosphodiesterase PDE4A5 by interaction with the immunophilin XAP2. J Biol
Buchbinder S, Bierhaus A, Zorn M, Nawroth PP, Humpert P, Schilling T. 2008. Aryl
hydrocarbon receptor interacting protein gene (AIP) mutations are rare in
patients with hormone secreting or non-secreting pituitary adenomas. Exp Clin
Endorcrinol Diabetes 116:625–628.
Cazabat L, Libe R, Perlemoine K, Rene-Corail F, Burnichon N, Gimenez-Roqueplo AP,
Dupasquier-Fediaevsky L, Bertagna X, Clauser E, Chanson P, Bertherat J,
Raffin-Sanson ML. 2007. Germline inactivating mutations of the aryl hydro-
carbon receptor-interacting protein gene in a large cohort of sporadic
acromegaly: mutations are found in a subset of young patients with
macroadenomas. Eur J Endocrinol 157:1–8.
Chahal HS, Chapple JP, Frohman LA, Grossman AB, Korbonits M. 2010. Clinical,
genetic and molecular characterisation of patients with familial isolated
pituitary adenomas. Trends Endocrinol Metab (in press).
Daly AF, Vanbellinghen JF, Khoo SK, Jaffrain-Rea ML, Naves LA, Guitelman MA,
Murat A, Emy P, Gimenez-Roqueplo AP, Tamburrano G, Raverot G, Barlier A,
De Herder W, Penfornis A, Ciccarelli E, Estour B, Lecomte P, Gatta B, Chabre O,
Sabate MI, Bertagna X, Garcia BN, Stalldecker G, Colao A, Ferolla P,
Wemeau JL, Caron P, Sadoul JL, Oneto A, Archambeaud F, Calender A,
Sinilnikova O, Montanana CF, Cavagnini F, Hana V, Solano A, Delettieres D,
Luccio-Camelo DC, Basso A, Rohmer V, Brue T, Bours V, Teh BT, Beckers A.
2007. Aryl hydrocarbon receptor-interacting protein gene mutations in familial
isolated pituitary adenomas: analysis in 73 families. J Clin Endocrinol Metab
D’Andrea LD, Regan L. 2003. TPR proteins: the versatile helix. Trends Biochem Sci
Das AK, Cohen PW, Barford D. 1998. The structure of the tetratricopeptide repeats of
protein phosphatase 5: implications for TPR-mediated protein-protein interac-
tions. EMBO J 17:1192–1199.
de Oliveira SK, Hoffmeister M, Gambaryan S, Muller-Esterl W, Guimaraes JA,
Smolenski AP. 2007. Phosphodiesterase 2A forms a complex with the co-chaperone
HUMAN MUTATION, Vol. 31, No. 8, 950–960, 2010
XAP2 and regulates nuclear translocation of the aryl hydrocarbon receptor. J Biol Download full-text
Dworakowska D, Wlodek E, Leontiou C, Igreja S, Cakir M, Teng M, Prodromou N,
Goth M, Grozinsky-Glasberg S, Gueorguiev M, Kola B, Korbonits M,
Grossman A. 2009. Activation of Raf/MEK/ERK and PI3K/Akt/mTOR pathways
in pituitary adenomas and their effects on downstream effectors. Endocr Relat
Farrell WE, Clayton RN. 2000. Molecular pathogenesis of pituitary tumors. Front
Gadelha MR, Prezant TR, Une KN, Glick RP, Moskal SF, Vaisman M, Melmed S,
Kineman RD, Frohman LA. 1999. Loss of heterozygosity on chromosome 11q13
in two families with acromegaly/gigantism is independent of mutations of the
multiple endocrine neoplasia type I gene. J Clin Endocrinol Metab 84:249–256.
Gadelha MR, Une KN, Rohde K, Vaisman M, Kineman RD, Frohman LA. 2000.
Isolated familial somatotropinomas: establishment of linkage to chromosome
11q13.1–11q13.3 and evidence for a potential second locus at chromosome
2p16–12. J Clin Endocrinol Metab 85:707–714.
Georgitsi M, De ME, Cannavo S, Makinen MJ, Tuppurainen K, Pauletto P, Curto L,
Weil RJ, Paschke R, Zielinski G, Wasik A, Lubinski J, Vahteristo P, Karhu A,
Aaltonen LA. 2008a. Aryl hydrocarbon receptor interacting protein (AIP) gene
mutation analysis in children and adolescents with sporadic pituitary adenomas.
Clin Endocrinol (Oxf) 69:621–627.
Georgitsi M, Heliovaara E, Paschke R, Kumar AV, Tischkowitz M, Vierimaa O,
Salmela P, Sane T, De ME, Cannavo S, Gundogdu S, Lucassen A, Izatt L,
Aylwin S, Bano G, Hodgson S, Koch CA, Karhu A, Aaltonen LA. 2008b. Large
genomic deletions of aryl hydrocarbon receptor interacting protein (AIP) gene
in pituitary adenoma predisposition. J Clin Endocrinol Metab 93:4146–4151.
Georgitsi M, Raitila A, Karhu A, Tuppurainen K, Makinen MJ, Vierimaa O,
Paschke R, Saeger W, van der Luijt RB, Sane T, Robledo M, De ME, Weil RJ,
Wasik A, Zielinski G, Lucewicz O, Lubinski J, Launonen V, Vahteristo P,
Aaltonen LA. 2007. Molecular diagnosis of pituitary adenoma predisposition
caused by aryl hydrocarbon receptor-interacting protein gene mutations. Proc
Natl Acad Sci USA 104:4101–4105.
Guarente L. 1983. Yeast promoters and lacZ fusions designed to study expression of
cloned genes in yeast. Methods Enzymol 101:181–191.
Hidalgo-de-Quintana J, Evans RJ, Cheetham ME, van der Spuy J. 2008. The Leber
congenital amaurosis protein AIPL1 functions as part of a chaperone
heterocomplex. Invest Ophthalmol Vis Sci 49:2878–2887.
Horvath A, Stratakis CA. 2008. Clinical and molecular genetics of acromegaly:
MEN1, Carney complex, McCune-Albright syndrome, familial acromegaly and
genetic defects in sporadic tumors. Rev Endocr Metab Disord 9:1–11.
Igreja SC, Chahal HS, Leontiou CA, Bolger GB, Chapple JP, Grossman AB,
Korbonits M. 2010. Functional characterisation of aryl hydrocarbon receptor
interacting protein (AIP) promoter and silent mutations. Endocr Abstr 21:P210.
Iwata T, Yamada S, Mizusawa N, Golam HM, Sano T, Yoshimoto K. 2007. The aryl
hydrocarbon receptor-interacting protein gene is rarely mutated in sporadic
GH-secreting adenomas. Clin Endocrinol (Oxf) 66:499–502.
Jaffrain-Rea ML, Angelini M, Gargano D, Tichomirowa MA, Daly AF, Vanbellinghen JF,
D’Innocenzo E, Barlier A, Giangaspero F, Esposito V, Ventura L, Arcella A,
Theodoropoulou M, Naves LA, Fajardo C, Zacharieva S, Rohmer V, Brue T,
Gulino A, Cantore G, Alesse E, Beckers A. 2009. Expression of aryl hydrocarbon
receptor (AHR) and AHR-interacting protein in pituitary adenomas: pathological
and clinical implications. Endocr Relat Cancer 16:1029–1043.
Jennings J, Georgitsi M, Holdaway I, Daly A, Tichomirowa M, Beckers A, Aaltonen L,
Karhu A, Cameron F. 2009. Aggressive pituitary adenomas occurring in young
patients in a large Polynesian kindred with a germline R271W mutation in the
AIP gene. Eur J Endocrinol 161:799–804.
Kelley LA, Sternberg MJ. 2009. Protein structure prediction on the Web: a case study
using the Phyre server. Nat Protoc 4:363–371.
Laenger A, Lang-Rollin I, Kozany C, Zschocke J, Zimmermann N, Ruegg J,
Holsboer F, Hausch F, Rein T. 2009. XAP2 inhibits glucocorticoid receptor
activity in mammalian cells. FEBS Lett 583:1493–1498.
Leontiou CA, Gueorguiev M, van der SJ, Quinton R, Lolli F, Hassan S, Chahal HS,
Igreja SC, Jordan S, Rowe J, Stolbrink M, Christian HC, Wray J, Bishop-Bailey D,
Berney DM, Wass JA, Popovic V, Ribeiro-Oliveira Jr A, Gadelha MR, Monson JP,
Akker SA, Davis JR, Clayton RN, Yoshimoto K, Iwata T, Matsuno A, Eguchi K,
Musat M, Flanagan D, Peters G, Bolger GB, Chapple JP, Frohman LA,
Grossman AB, Korbonits M. 2008. The role of the AIP gene in familial and
sporadic pituitary adenomas. J Clin Endocrinol Metab 93:2390–2401.
Lin BC, Sullivan R, Lee Y, Moran S, Glover E, Bradfield CA. 2007. Deletion of the aryl
hydrocarbon receptor-associated protein 9 leads to cardiac malformation and
embryonic lethality. J Biol Chem 282:35924–35932.
Matsuno A, Teramoto A, Yamada S, Kitanaka S, Tanaka T, Sanno N, Osamura RY,
Kirino T. 1994. Gigantism in sibling unrelated to multiple endocrine neoplasia:
case report. Neurosurgery 35:952–955.
McCarthy MI, Noonan K, Wass JA, Monson JP. 1990. Familial acromegaly: studies in
three families. Clin Endocrinol (Oxf) 32:719–728.
Melmed S. 2006. Medical progress: acromegaly. N Eng J Med 355:2558–2573.
Montanana CF, Daly AF, Tichomirowa MA, Vanbellinghen JF, Jaffrain-Rea ML,
Trescoli Serrano C, Riesgo Suares P, Gomez Vela J, Tenes S, Bours V, Beckers A.
2009. TSH-secreting pituitary adenoma in a male patient with a novel missense
AIP mutation. Proc of the 91st Annual Meet of the Endocrine Soc; P1-668.
Oesch-Bartlomowicz B, Huelster A, Wiss O, ntoniou-Lipfert P, Dietrich C, Arand M,
Weiss C, Bockamp E, Oesch F. 2005. Aryl hydrocarbon receptor activation by
cAMP vs. dioxin: divergent signaling pathways. Proc Natl Acad Sci USA
Ogino S, Wilson RB, Gold B, Flodman P. 2007. Bayesian risk assessment in genetic
testing for autosomal dominant disorders with age-dependent penetrance.
J Genet Couns 16:29–39.
Pestell RG, Alford FP, Best JD. 1989. Familial acromegaly. Acta Endocrinol (Copenh)
Petrulis JR, Perdew GH. 2002. The role of chaperone proteins in the aryl hydrocarbon
receptor core complex. Chem Biol Interact 141:25–40.
Raitila A, Georgitsi M, Karhu A, Tuppurainen K, Makinen MJ, Birkenkamp-
Demtroder K, Salmenkivi K, Orntoft TF, Arola J, Launonen V, Vahteristo P,
Aaltonen LA. 2007. No evidence of somatic aryl hydrocarbon receptor
interacting protein mutations in sporadic endocrine neoplasia. Endocr Relat
Rohlff C, Ahmad S, Borellini F, Lei J, Glazer RI. 1997. Modulation of trans-
cription factor Sp1 by cAMP-dependent protein kinase. J Biol Chem 272:
Sikorski RS, Boguski MS, Goebl M, Hieter P. 1990. A repeating amino acid motif in
CDC23 defines a family of proteins and a new relationship among genes
required for mitosis and RNA synthesis. Cell 60:307–317.
Soares BS, Eguchi K, Frohman LA. 2005. Tumor deletion mapping on chromosome
11q13 in eight families with isolated familial somatotropinoma and in 15
sporadic somatotropinomas. J Clin Endocrinol Metab 90:6580–6587.
Tahir A, Chahal HS, Korbonits M. 2010. Molecular genetics of the AIP gene in
familial pituitary tumorigenesis. Prog Brain Res (in press).
Toledo RA, Lourenco Jr DM, Liberman B, Cunha-Neto MB, Cavalcanti MG,
Moyses CB, Toledo SP, Dahia PL. 2007. Germline mutation in the aryl
hydrocarbon receptor interacting protein gene in familial somatotropinoma.
J Clin Endocrinol Metab 92:1934–1937.
Vargiolu M, Fusco D, Kurelac I, Dirnberger D, Baumeister R, Morra I, Melcarne A,
Rimondini R, Romeo G, Bonora E. 2009. The tyrosine kinase receptor RET
interacts in vivo with aryl hydrocarbon receptor-interacting protein to alter
survivin availability. J Clin Endocrinol Metab 94:2571–2578.
Vierimaa O, Georgitsi M, Lehtonen R, Vahteristo P, Kokko A, Raitila A,
Tuppurainen K, Ebeling TM, Salmela PI, Paschke R, Gundogdu S, De ME,
Makinen MJ, Launonen V, Karhu A, Aaltonen LA. 2006. Pituitary adenoma
predisposition caused by germline mutations in the AIP gene. Science
Yamada S, Yoshimoto K, Sano T, Takada K, Itakura M, Usui M, Teramoto A. 1997.
Inactivation of the tumor suppressor gene on 11q13 in brothers with familial
acrogigantism without multiple endocrine neoplasia type 1. J Clin Endocrinol
Yaneva M, Daly AF, Tichomirowa MA, Vanbellinghen JF, Hagelstein MT, Bours V,
Zacharieva S, Beckers A. 2008. Aryl hydrocarbon receptor interacting protein
gene mutations in bulgarian FIPA and young sporadic pituitary adenoma
patients. Proc of the 90th Annual Meet of the Endocrine Soc:P3-520.
HUMAN MUTATION, Vol. 31, No. 8, 950–960, 2010