Department Human Molecular Genetics(Hans-Hilger Ropers)

Introduction

In search of clinically relevant genetic risk factors for common disorders, genome-wide association studies (GWAS) have been performed since more than

15 years, with meagre results. Now even the most prominent advocate of this research direction has thrown the sponge1 and world-wide, rare disorders have come into focus of genome research again2. We were among the first to point out the inherent difficulties of GWAS in complex diseases, including their genetic heterogeneity, and to stress the importance of studying single gene disorders as a better alternative. Consequently, the research of our department has revolved around monogenic disorders since its inception3.

During the past decade, we have increasingly focused on intellectual disability (ID) and related disorders. ID, also called mental retardation or early-onset cognitive impairment, is the biggest unsolved problem of clinical genetics and a far heavier socio-economic burden than, e.g., cancer4. Of the several thousand gene defects that may give rise to ID, only a few hundred have been identified so far, since many forms of ID are clinically indistinguishable and because in Western Societies, families tend to be small. We circumvented this problem by forging international cooperations, e.g. with a potent group in Iran. This enabled us to study familial forms of ID in a systematic manner, and put us in an excellent position for scaling up this research when high-throughput sequencing techniques became available, as outlined below.

This research would not have been possible without substantial financial support from the Max Planck Society and additional funding from the German Federal Ministry of Education and Research. Both came to an end in October 2011, when the Head of the Department reached his official retirement age, but was re-installed as Acting Director for a period of three years. Since then, a deconstruction plan is in place, which entails a progressive reduction of the structural budget until the end of October 2014, when the department will be closed. However, additional financial support has been obtained through an EU grant (Genetic and Epigenetic Networks in COgnitive DISorders [GENCODYS], FP7 reference no. 241995,

05/2010 – 04/2015) which will partially compensate the diminishing structural resources.

In accordance with the deconstruction plan, the groups of Tim Hucho (Signal Transduction in Pain and Mental Retardation) and Diego Walther (Monoamine Signalling and Disease, Mouse Lab.) were discontinued when the contracts of their leaders expired. Andreas Tzschach, our former clinical geneticist, left for the University of Tuebingen in 2011, and his position was filled by Thomas Wienker, a human geneticist and retired professor of biostatistics from the University of Bonn. Andreas Kuss was appointed as a professor (W2) by the University of Greifswald, and since 2011, Luciana Musante, a former post-doctoral fellow in the group of Vera Kalscheuer, is now in charge of our research into recessive ID. While we are still maintaining close ties with Wei Chen, now at the Max Delbrück Center in Berlin, his part-time appointment at the MPIMG was discontinued in 2011, and his former post-doctoral fellow Hao Hu took over his task as the bioinformatician of our department. Finally, the part-time appointment of Susann Schweiger (University of Dundee and future Head of Human Genetics at the University of Mainz) has also been terminated in 2011.

In view of the reduced size of the department, which is also due to the fact that we can no longer hire PhD students, most former groups have lost their critical mass. At the same time, the research of the remaining scientists converged and their collaboration intensified. Therefore, their scientific achievements are no longer presented separately, except for Tim Hucho, recently appointed as W2 professor at the University of Cologne, and Reinhard Ullmann’s group with its gradually diverging orientation and own DFG support.

Scientific methods and findings

Next generation sequencing revolutionizes the identification of ID genes

(Wei Chen, Hao Hu, Vera Kalscheuer, Reinhard Ullmann, Andreas Tzschach, Andreas Kuss)

To elucidate the genetic defects underlying ID and related disorders, we have employed a wide spectrum of approaches, including breakpoint mapping in patients with disease-associated balanced chromosome rearrangements (DBCRs), screening for disease-associated copy number variants by array CGH, and linkage mapping in families and mutation screening of candidate genes, as outlined previously3. While array CGH-based screening for pathogenic copy number variants, the study of DBCRs and linkage mapping in patients and families remain useful strategies for the elucidation of genetic disorders, as illustrated by several recent publications of our group, our decision paid off to invest early into genomic enrichment techniques, high throughput sequencing and the storage, handling and interpretation of next generation sequencing data.

Development of mutation detection pipelines

(Hao Hu; together with Stefan Haas and co-workers, Dept. Computational Molecular Biology)

Various members of the Department of Computational Molecular Biology (Head: Martin Vingron) contributed to this effort by developing a bioinformatic mutation detection pipeline, which was first used to look for de novo mutations on the X-chromosome in patient-parent trios with a suspected X-linked dominant disorders (Chen W. et al, unpublished observations). Later on, this pipeline was instrumental in our comprehensive collaborative effort to identify the molecular defects underlying X-linked mental retardation (see below).

Independently, Hao Hu developed another algorithm for identifying pathogenic changes in whole genome and whole exome sequences. This algorithm has been employed successfully to look for mutations in consanguineous families with autosomal recessive ID and has been described in several publications5; a more comprehensive description is in preparation. Since 2010, these methods have become the mainstay of our research into the genetic causes of ID and related disorders.

X-linked ID genes: draining the pond

(Vera Kalscheuer, Hao Hu, Chen Wei, Thomas Wienker; in cooperation with Stefan Haas, Tomasz Zemojtel, Martin Vingron, Dept. Computational Molecular Biology)

Employing a custom-made hybrid capture kit to enrich 7591 X-chromosomal exons, or 875 genes, we have performed targeted exon sequencing in 248

European families with X-linked forms of ID. In the vast majority of these families, X-linkage was virtually certain because of affected males in separate sibships that were connected through healthy females. Apparently deleterious DNA variants were identified in 13 genes that had not been implicated in ID before, and their identity as novel genes for X-linked ID (XLID) was corroborated in various ways. This study raises the number of known XLID genes to 110. Using the same parameters to distinguish pathogenic from clinically irrelevant sequence variants, we have also reanalyzed the results of a previous study encompassing 208 Caucasian families, which had been screened for mutations by large-scale Sanger sequencing6. Under the (plausible) assumption that the cohorts analyzed by the two studies are part of the same population, this enabled us to estimate the total number of XLID genes as 123 (95% confidence limits: 91-155). This estimate is lower than expected and cannot be reconciled easily with our finding that mutations in the known 110 genes account for at most 71% of the XLID families. One possible explanation for this discrepancy is that most of the missing mutations may reside in non-coding, e.g. intronic sequences which were not analyzed in these studies (Kalscheuer et al, unpublished).

A plethora of novel genes for autosomal recessive forms of ID (ARID)

(Andreas Kuss, Andreas Tzschach, Hao Hu, Masoud Garshasbi, Luciana Musante, Thomas Wienker)

Following up on a previous, pioneering study to identify novel ARID loci and to assess the genetic heterogeneity of this condition3, we have performed array-based SNP typing and linkage mapping in 300 consanguineous Iranian and German families. In 27 of these families, a single homozygous interval was observed, and at least 14 novel ARID loci could be identified7. Starting in 2006, when only three ARID genes had been reported,3 mutation screening of all genes located single linkage intervals has revealed numerous novel genes for syndromic or non- syndromic forms of ID (see Table 1).

Gene

Location

Function

Ethnicity

Reference

GRIK2

6q16.3

Involved in the transmission of light signals from the retina to the hypothalamus, Involved in the maturation of microcircuits and network formation in brain areas

Iranian

Motazacker MM et al., Am J Hum Genet 2007;

81: 792–798

TUSC3

8p22

Putative Mg2+ transporter, required for cellular Mg2+ uptake. Indispensable for normal vertebrate embryonic development.

Iranian, French

Garshasbi M et al., Am

J Hum Genet 2008; 82:

1158–1164

VLDLR

9p24

Part of the reelin signaling pathway, which is involved in neuroblast migration in the cerebral cortex and cerebellum

Iranian, Canadian, Turkish

Abbasi Moheb L et al., Euro J Hum Genet 2008;

16: 270–273

TRAPPC9

8q24.3

Enhancer of the cytokine-induced NF- (kappa)B signaling pathway, having an essential function in post mitotic neurons as opposed to neural progenitors

Iranian, Pakistani, Tunisian, Israeli

Mir A et al., Am J Hum

Genet 2009; 85: 909-915

SRD5A3

4q12

Polyprenol reductase with a crucial role in N-linked protein glycosylation that

is required for converting polyprenol to dolichol.

Iranian, Emirati, Turkish, Polish

Kahrizi K et al., Euro

J Hum Genet 2011;

19:115–117

ZC3H14

14q31.3

May contribute to control of gene expression in human cells through binding poly(A) RNA

Iranian

Pak CH et al., PNAS

2011; 108:12390-95

ST3GAL3

1p34.1

Transfers sialic acid to terminal positions on the carbohydrate groups of glycoproteins and glycolipids that are

key determinants for a variety of cellular recognition processes

Iranian

Hu H et al, Am J Hum

Genet 2011; 89:407-414

NSUN2

5p15.31

RNA methyltransferase that methylates tRNAs, and possibly RNA polymerase III transcripts. May act downstream of Myc to regulate epidermal cell growth and proliferation

Iranian

Abbasi Moheb L et al., Am J Hum Genet 2012;

90: 847-55

ZNF526

19q13.2

Involved in transcriptional regulation, role in regulation of translation

Iranian

Abbasi Moheb L et al., ESHG meeting 2011

Table 1: Novel molecular defects underlying syndromic and non-syndromic ARID.

In 2010, by combining targeted exon enrichment and next generation sequencing, we extended these studies to 136 consanguineous ARID families with more than one linkage interval. In 78 of these, a single plausible causative mutation could be identified, involving 22 known and 50 novel candidate genes. This study, published in 2011,5 quadruplicated the number of (candidate) genes for non- syndromic forms of ARID, which were found to be more common than syndromic forms. For the vast majority of these genes, pathogenic mutations were only seen in a single family. This corroborates our previous observations and suggests that none of these gene defects can account for more than a few percent of all forms of ID in Iran – but it does not exclude founder mutations for ARID in one or several of the 7 or 8 Iranian sub-populations.

Currently, we reanalyze families with more than one plausible mutation, e.g. by studying knock-down fly models for the relevant gene defects (see below), as well as families where no single homozygous mutation has been found. In some of these families, the ID may be due to compound heterozygosity, i.e., it may be unrelated to the parental consanguinity, or the causative mutations may hide in introns or other non-coding sequences that have not been investigated so far. To unveil these missing mutations, whole genome sequencing has been performed in

11 of these families, and analysis of the results is ongoing.

In parallel, we have embarked on a second, even larger study encompassing almost 150 consanguineous families from Iran and Germany, including all remaining large families collected by our Iranian partner in the course of this long-standing collaboration. In most of these families, SNP typing revealed multiple homozygous linkage intervals, which renders targeted exon enrichment with custom-made arrays costly and tedious. Therefore, and in order to detect compound heterozygosity, we have turned to whole exome sequencing instead, Currently, we reanalyze families with more than one plausible mutation, e.g. by studying knock-down fly models for the relevant gene defects (see below), as well as families where no single homozygous mutation has been found. In some of these families, the ID may be due to compound heterozygosity, i.e., it may be unrelated to the parental consanguinity, or the causative mutations may hide in introns or other non-coding sequences that have not been investigated so far. To unveil these missing mutations, whole genome sequencing has been performed in

11 of these families, and analysis of the results is ongoing.

Novel ARID genes encode components of key cellular protein networks (from Najmabadi et al., 2011).

Novel ARID genes encode components of key cellular protein networks (from Najmabadi et al., 2011).

Diagnostic aspects

(Thomas Wienker, Wei Chen, Hao Hu; Thomasz Zemojtel, Dept. Computational Molecular Biology)

High-throughput sequencing techniques have not only revolutionized the elucidation of single gene disorders, but also provided the basis for comprehensive diagnostic tests to rule out mutations in all known disease genes. In collaboration with Stephen Kingsmore (Children’s Mercy Hospital, Kansas City, USA), the Pediatric Department of the Berlin University Hospital Charité and Wei Chen at the MDC Berlin, we have developed a clinical entry test for children with severe ID and/or unexplained developmental delay. This test, baptized ‘MPIMG1’, encompasses numerous severe childhood disorders, all published ID genes as well as the many novel ones identified by our group. In principle, this test can also be employed for non-invasive preconception carrier detection to rule out elevated parental risks for children with severe recessive disorders. This application renders it particularly useful for consanguineous parents and for countries from the so-called ‘Consanguinity Belt’ that extends from the Maghreb to India.

Outlook

Until our department will be officially closed in late fall 2014 and the EU-FP7 project will expire in April 2015, most of our remaining resources will be used to successfully conclude three ongoing, closely related projects.

First, the long-standing collaboration with our Iranian partner will reach its natural end when all ARID families collected since 2004 have been analyzed and funding of both groups by the afore-mentioned EU project will be discontinued. We expect that until then, our research into autosomal and X-linked forms of ID will remain internationally competitive.

Secondly, provided that residual administrative hurdles can be overcome, we will implement our novel MPIMG1 test at the Charité – Universitätsmedizin Berlin, the Children’s Mercy Hospital in Kansas City and elsewhere as a first-line diagnostic tool for all known genetic defects that cannot be readily diagnosed by clinical examination alone. Thereafter, we intend to offer this diagnostic and carrier test to improve genetic health care in selected countries with frequent parental consanguinity, intellectual disability and developmental delay.

Finally, in collaboration with a Dutch group, we are about to shed light on the old question whether ID genes also have a role as determinants of the normal IQ distribution, as postulated by Lehrke9 40 years ago. Targeted exon enrichment and next generation sequencing in ~ 170 selected ID genes will be performed to test the hypothesis that there is an inverse relationship between the IQ of the proband and the number of subtle mutations in these genes. While the analysis of these data may turn out to be quite difficult and time-consuming, we expect that this investigation will be completed well within the available time-frame, even if it should turn out that follow-up studies will be required to answer this question in full.