„The Biological Role of Fetuin-B in Reproductive Biology and Fetuin-B as Target for Contraception“

Von der Fakultät für Mathematik, Informatik und Naturwissenschaften der RWTH Aachen University zur Erlangung des akademischen Grades einer Doktorin der Naturwissenschaften genehmigte Dissertation

vorgelegt von

M. Sc. Julia Elisabeth Flöhr aus Aachen

Berichter: Univ.-Prof. Dr. rer. nat. Willi Jahnen-Dechent Univ.-Prof. Dr. rer. nat. Marc Spehr

Tag der mündlichen Prüfung: 21.10.2016

Diese Dissertation ist auf den Internetseiten der Universitätsbibliothek online verfügbar.

2

Parts of this thesis have been published in peer-reviewed articles:

Floehr J, Dietzel E, Neulen J, Rösing B, Weissenborn U, Jahnen-Dechent W. Association of high fetuin-B concentrations in serum with fertilization success rate in IVF: a cross-sectional pilot study. Hum Reprod. 2016;31:630–7.

Floehr J, Dietzel E, Schmitz C, Chappell A, Jahnen-Dechent W. Down-regulation of the -derived plasma fetin-B mediates reversible female infertility. Mol Hum Reprod. 2016. [Epub ahead of print].

Dietzel E, Floehr J, Jahnen-Dechent W. The biological role of fetuin-B in female reproduction. Ann Reprod Med Treat. 2016; 1:1003.

3

4

Table of content

Summary 7

1 Introduction 11 1.1 Fetuin family 11 1.1.1 Physiological role of fetuin-B in zona pellucida hardening 12 1.2 Female (in)fertility 15 1.2.1 Oogenesis 16 1.2.2 Hormonal regulation of oogenesis 18 1.3 Contraception 21 1.4 Antisense oligonucleotides – chemistry and mechanism 22 1.5 The mouse as model organism in reproductive biology 24 1.6 The aim of the study 26

2 Experimental Procedures 27 2.1 Fetuin-B in human reproductive biology 27 2.1.1 Ethical approval 27 2.1.2 Study population and design 27 2.1.3 Serum and follicular fluid sampling 28 2.1.4 Protein and hormone measurements 28 2.1.5 Ovastacin inhibition assay 28 2.1.6 Statistical analysis 29 2.2 Molecular fetuin-B probes - Monoclonal antibody production 30 2.2.1 treatment 30 2.2.2 Production and purification of recombinant mouse fetuin-B 30 2.2.3 Immunization of fetuin-B deficient mice 31 2.2.4 Fusion of hybridoma cells 32 2.2.5 Specificity of hybridoma cell-derived monoclonal antibodies determined by ELISA 33 2.2.6 Selection of stable antibody-producing cells 33 2.2.7 Limiting dilution-cloning 34 2.2.8 Cryopreservation of hybridoma cells 34 2.2.9 Specificity of hybridoma cell-derived monoclonal antibodies determined by Western blot 35 2.2.10 Immunoprecipitation of serum fetuin-B by monoclonal antibodies 35 2.3 Molecular fetuin-B probes - Fetuin-B as target for contraception 36 2.3.1 Generation of fetuin-B/ovastacin double deficient mice 36 2.3.2 Fetuin-B and Ovastacin genotyping 36 2.3.3 In vitro fertilization 38 2.3.4 Antisense oligonucleotides chemistry and synthesis 40 2.3.5 Serum fetuin-B down-regulation by ASO treatment 41 2.3.6 Serum fetuin-B determination 41 2.3.7 Mating of fetuin-B ASO-treated females 41 2.3.8 Offspring of fetuin-B ASO-treated females 41 2.3.9 Fetub and Gapdh RNA quantification 42 2.3.10 Serum chemistry 42 2.3.11 Statistical analysis 43

3 Results 45 3.1 Fetuin-B in human reproductive biology 45 3.1.1 Serum fetuin-B in 45 Table of content

3.1.2 Fetuin-B expression is stimulated by ethinyl estradiol 45 3.1.3 Fetuin-B levels in serum and follicular fluid are tightly associated 47 3.1.4 Fetuin-B expression is stimulated by high endogenous estradiol during ovarian stimulation and during pregnancy 49 3.1.5 Serum fetuin-B is associated with fertilization rate in IVF 51 3.1.6 Ovastacin is inhibited by human fetuin-B 52 3.2 Molecular fetuin-B probes - Monoclonal antibody production 53 3.2.1 Purification of recombinant mouse fetuin-B 53 3.2.2 Monoclonal fetuin-B antibody production and evaluation 54 3.3 Molecular fetuin-B probes – Fetuin-B as target for contraception 64 3.3.1 Confirming fetuin-B as a target for contraception in vivo - Fertility recovery of Fetub deficient mice in Fetub/Astl double deficient mice 64 3.3.2 Dose-finding study of fetuin-B ASO 66 3.3.3 Fetuin-B ASO mediated down-regulation causes infertility 69 3.3.4 Contraceptive effect of fetuin-B ASO is reversible 76 3.3.5 Permissive serum fetuin-B range required for contraception 79

4 Discussion 83 4.1 Fetuin-B in human reproductive biology 83 4.2 Molecular fetuin-B probes 86 4.2.1 Monoclonal antibody production 86 4.2.2 Fetuin-B as target for contraception 87

5 Conclusion and Future Aspects 95

References 97

Abbreviations 111

Figures 115

Tables 117

Acknowledgment 119

6

Summary

Summary In mice fetuin-B is essential for fertilization. Fetuin-B deficiency (Fetub-/-) leads to female infertility due to premature zona pellucida hardening. The zona pellucida (ZP) is a glycoprotein matrix surrounding the oocyte. The penetration of the ZP by sperm is a critical factor for fertilization; fetuin-B maintains the fertility of oocytes by inhibiting premature ZP. The fetuin-B is well conserved between mammals with 61% sequence homology in mice and humans. To study the role of fetuin-B in human female reproduction serum and follicular fluid fetuin-B were determined in healthy volunteers and in patients consulting the fertility clinic. The results suggest an estrogen-mediated regulation of hepatic fetuin-B expression. Furthermore, serum fetuin-B increased in women with successful in vitro fertilization (IVF) while serum fetuin-B remained unchanged in women with fertilization failure. Using fetuin-B as an additional marker in assisted reproductive technology may increase the fertilization rate. Biochemical studies showed that fetuin-B is a potent ovastacin inhibitor. Ovastacin induces ZP hardening by cleavage of the zona pellucida protein 2 (ZP2). Using conventional breeding, fetuin-B/ovastacin double deficient (Fetub-/-, Astl-/-) mice were generated. Fertility of female Fetub-/-, Astl-/- mice showed in vivo that fertility of Fetub-/- females was restored by additional ovastacin deficiency. Fetub-/-, Astl-/- females produced healthy offspring, confirming the proteinase ovastacin as the target of the inhibitor fetuin-B. Following natural mating, the litter size of Fetub-/-, Astl-/- females were highly variable from unusually low to normal litter sizes. Because the IVF rates were unusually high in these oocytes a defect in fertilization was definitely not the primary reason for small litter sizes suggesting that embryo abortion occurred following successful fertilization. This was attributed to a complete loss of fertilization-induced ZP hardening, and thus a lack of pre-implantation embryo stability. Nevertheless, the fact that the infertility of Fetub-/- mice was fully restored in Fetub-/-, Astl-/- mice underscored the important role of fetuin-B in fertilization, which might be exploited as a potential contraceptive target. Corroborating this view, antisense oligonucleotide-mediated down-regulation of fetuin-B expression reversibly prevented pregnancy. Due to the high homology between fetuin-B and ovastacin in mice and humans it is hypothesized that pharmacological fetuin-B down-regulation may be also contraceptive in women.

7

8

Zusammenfassung

Zusammenfassung Das Plasmaprotein Fetuin-B ist essenziell für die Befruchtung. In weiblichen Fetuin-B defizienten Mäusen (Fetub-/-) härtet die Zona Pellucida frühzeitig, so dass es nicht zur Verschmelzung von Spermium und Eizelle kommt. Die Zona Pellucida (ZP) ist eine Glykoproteinmatrix, die die Eizelle umgibt und ihre Penetration ein kritischer Punkt der Befruchtung. In Wildtyp Mäusen verhindert Fetuin-B eine frühzeitige Härtung der ZP und erhält somit die Befruchtungsfähigkeit der Eizelle. In Säugern ist das Fetuin-B Gen gut konserviert und weist zwischen Mensch und Maus eine Sequenzhomologie von 61% auf. Um die Funktion von Fetuin-B in der weiblichen Reproduktion zu untersuchen, wurde die Fetuin-B Konzentration in Serum und Follikelflüssigkeit von gesunden Probanden/Probandinnen und in Patientinnen der Fertilitätsklinik bestimmt. Die Ergebnisse lassen auf eine Estrogen vermittelte hepatische Regulation von Fetuin-B schließen. Des Weiteren konnte gezeigt werden, dass die Serum Fetuin-B Konzentration im Vorfeld einer erfolgreichen in vitro Fertilisation (IVF) ansteigt, während es bei Frauen, deren IVF ohne Befruchtungserfolg verlief, konstant blieb. Um zukünftig die Befruchtungsrate zu steigern, könnte es somit hilfreich sein Fetuin-B als einen zusätzlichen Marker in der assistierten Reproduktionsbiologie zu verwenden. Durch frühere Studien ist bekannt, dass Fetuin-B ein potenter Inhibitor der Proteinase Ovastacin ist. Ovastacin induziert die Härtung der ZP durch Spaltung des Zona Pellucida 2 (ZP2). Mithilfe einer Verpaarungsstudie wurden Fetuin-B/Ovastacin doppelt defiziente (Fetub-/-, Astl-/-) Mäuse erzeugt und gezeigt, dass die Infertilität der Fetub-/- Weibchen durch eine zusätzlich eingebrachte Ovastacin Defizienz aufgehoben werden konnte. Durch die Geburt von gesunden Nachkommen konnte somit die Proteinase Ovastacin als Target des Inhibitors Fetuin-B bestätigt werden. Eine große Variabilität in der Wurfgröße nach natürlicher Verpaarung der Fetub-/-, Astl-/- Mäuse und die gleichzeitig hohe Befruchtungsrate nach IVF lassen vermuten, dass die Oozyten zwar sehr gut befruchtungsfähig sind, sich anschließend aber nicht weiterentwickeln. Der Grund hierfür liegt vermutlich im Ausbleiben der befruchtungsinduzierten Härtung der ZP, infolge dessen der Präimplantationsembryo an Stabilität verliert. Die wieder hergestellte Fertilität der Fetub-/-, Astl-/- Weibchen hob Fetuin-B als potenzielles kontrazeptives Target hervor. Tatsächlich konnte Fetuin-B durch die Behandlung mit Antisense Oligonukleotiden herunterreguliert und somit Schwangerschaften verhindert werden. Nach Beendigung der Behandlung zeigte sich, dass der kontrazeptive Effekt reversibel war. Die große Homologie von Fetuin-B und Ovastacin zwischen Mensch und 9

Zusammenfassung

Maus lässt vermuten, dass das Target Fetuin-B auch bei Frauen als Verhütungsmittel Bedeutung finden könnte.

10

Introduction

1 Introduction

1.1 Fetuin family In 1944, fetuin was first described by Pedersen 1. Fetuin, derived from the Latin word fetus, took its name as most abundant plasma protein in fetal calf serum. Fetuin-B (genetic symbol FETUB) belongs, as well as fetuin-A/alpha-2-Heremans Schmid-glycoprotein (AHSG), kininogen (KNG) and histidine-rich glycoprotein (HRG) to the type 3 members of the cystatin superfamily of cysteine proteinase inhibitors 2. These plasma proteins have a characteristic and conserved structure. They are multidomain glycoproteins with disulfide bridges in distinct distances and have at least two aminoterminal cystatin-like domains (Fig. 1). The proteins are predominately expressed in the liver wherefrom they are secreted into the circulation 3.

Figure 1. Structural comparison of type 3 cystatins. Fetuin-A (AHSG), fetuin-B (FETUB), histidine-rich glycoprotein (HRG) and kininogen (KNG), subdivided in low molecular weight kininogen (LK) and high molecular weight kininogen (HK), are type 3 members of the cystatin superfamily. Characteristically they present at least two aminoterminal cystatin domains with disulfide bridges in distinct distances indicated by the black lines above each domain structure. The scale bar below states the amino acid length. Numbers at the left represent the length of the leader sequence, numbers at the right the amino acid length in total. (modified after 168)

Fetuins thus have homologous cystatin protein domains, a conserved disulfide bridge pattern, and a signature amino acid sequence called “fetuin motif” (LETXCHXLDPTP) 4. Fetuin-B has a truncated form, LETGCHVL (Fig. 4). Fetuin-A and fetuin-B are paralogous in direct proximity on the chromosomal level. They developed by an evolutionary duplication event. In the human , fetuin-A and fetuin-B are located on , in mice on 11

Introduction chromosome 16. As mentioned before both proteins have two cystatin-like domains (D1 and D2) and a third carboxyl terminal domain (D3) 2. The structural analogy continues in a similar amino acid length. In mice fetuin-A has 345 and fetuin-B 388 amino acids; in humans 367 and 382 amino acids, respectively. All coding regions of fetuin proteins comprise seven exons, but murine fetuin-B transcription starts in an additional exon that does however not contain any translated nucleotides 3,5–7. Albeit the basic structure is similar, differences in fetuin-A and fetuin-B have been shown. Fetuin-A contains a calcium binding side in domain 1 that is absent in fetuin-B 4. Further it was described by Olivier and colleagues that fetuin-B has two proteinase inhibitory sequences referred to as Kunitz motifs: only one of the same is contained in fetuin-A 4. Whether or not these functional motifs are active as proteinase inhibitors is unclear. In 2003, it was shown that the tissue distribution varies slightly between both fetuins. In humans, both fetuins are predominately expressed in the liver, and low level fetuin-B expression was also reported in the placenta. Like humans, mice likewise produce most fetuin-A and fetuin-B in the liver. Additionally, whole embryo, and tongue, and ovary all showed low level fetuin-B expression 3. Human and murine fetuin-A and fetuin-B proteins were detected at ~60 kDa on a sodium dodecyl sulfate polyacrylamide gel (SDS- PAGE). According to the number of amino acids, fetuin-A has a slightly lower molecular weight on SDS-PAGE than fetuin-B. However, due to glycosylation and further post- translational modifications the apparent molecular mass detected on SDS-PAGE is higher than the theoretical molecular weight. The protein molecular weight without post-translational modifications of fetuin-A is 37 kDa – 39 kDa, for fetuin-B 42 kDa. Despite the striking structural similarities and the fact that both fetuins are secreted plasma proteins, they have very different physiological roles 8–10. While fetuin-A plays a key role in mineralized matrix metabolism and has anti-inflammatory effects 10–14, fetuin-B is essential for fertilization 8.

1.1.1 Physiological role of fetuin-B in zona pellucida hardening To study the role of fetuin-B in vivo, fetuin-B deficient (Fetub-/-) mice were generated. The fetuin-B gene was deleted by homologous recombination, following an established protocol from Hogan and coworkers in C57BL/6N mice and by conventional backcrossing in DBA/2 mice 8,15,16. The mice showed an inconspicuous appearance with regards to gross morphology, blood pressure, glucose tolerance, weight and histology of major organs 16. Fetub-/- mice were morphological and physiological normal and had comparable serum parameters to wildtype 12

Introduction mice. By a mating study it was revealed that Fetub-/- female mice are completely infertile. Female fetuin-B hemizygous (Fetub+/-) and male Fetub-/- and Fetub+/- mice were not affected. This phenotype is strain independent. In 2013, it was proved that infertility of female Fetub-/- mice, C57BL/6N and DBA/2, was caused by premature zona pellucida (ZP) hardening (Fig. 2) 8.

Figure 2. Physiological role of fetuin-B in fertilization. In wildtype mice fetuin-B inhibits ovastacin release during egg maturation (top, left). After fertilization (top right) ovastacin overwhelms fetuin-B concentration and fertilization-induced zona pellucida hardening occurs. In fetuin-B deficient mice (bottom) precocious release of ovastacin by the cortical granules leads to premature zona pellucida hardening. Hence no sperm is able to penetrate the oocyte. Female fetuin-B deficient mice are infertile. (modified after 8)

The ZP is a glycoprotein matrix surrounding the oocyte. ZP proteins control sperm attachment, penetration and thus fertilization rate. The human ZP contains four zona proteins (ZP1 - ZP4) 17–19, the mouse ZP is comprised of three proteins (ZP1 - ZP3) 20. Mice carry a pseudogene for ZP4 18. Current knowledge suggests that the mouse ZP composed of ZP2 and ZP3 forming heterodimers cross-linked by homodimers of ZP1 (Fig. 3) 21,22.

Figure 3. Scheme of the murine zona pellucida consisting of three zona pellucida proteins (ZP). In a ratio of 1:1 ZP2 and ZP3 build filamentous structures connected by ZP1 homodimers. (modified after 21)

13

Introduction

Physiologically, ZP hardening occurs after fertilization, and involves the cleavage of ZP proteins by proteinases released from the cortical granules of oocytes, an event triggered by sperm entry (Fig. 2) 23,24. The metalloproteinase ovastacin, cleaves ZP2 to the fragmented ZP2f, and thus triggers definitive ZP hardening 25. ZP hardening can be monitored by analyzing the cleavage of ZP2 19,24, the sperm attachment 23 as well as the resistance of the ZP to enzymatic digestion employing various experimental protocols 26,27. Premature ZP hardening, the hardening of the ZP before fertilization, is a hitherto unspecified complication of in vitro fertilization (IVF) both in humans (International Statistical Classification of Diseases and Related Health Problems, ICD N98.9) and in mice, which drastically reduces fertilization rate 28,29. Published data suggest that the microscopic appearance of the ZP is a useful marker for successful pregnancies following IVF. In humans, the ZP thickness is associated with the fertilization rate and is generally believed to indicate the state of ZP hardening 30. The ZP structure is highly conserved in mammals rendering mice an appropriate model to study the role of ZP components in mammalian fertilization 31. In mice, premature ZP hardening occurs in vivo, if fetuin-B is absent (Fig. 2) 8. Fetuin-B deficient females have reduced ZP thickness, prolonged ZP digestion time by chymotrypsin, and reduced sperm attachment 8. Cleavage of ZP2 into ZP2f was demonstrated in post-ovulated oocytes while preovulatory oocytes contained predominantly uncleaved ZP2 protein, suggesting that the hardening process occurred at the time of ovulation. This is consistent with the observation that cortical granules exocytose already before fertilization and thereby the ZP2 proteinase ovastacin is released 25,32. Quantitative analysis revealed cortical granule loss by one third from the stage of mature to ovulated oocytes 33. Furthermore, in vitro studies demonstrated that fetuin-B is a potent ovastacin inhibitor with an IC50 (half maximal inhibitory 8 concentration) of 75 nM . As a consequence of premature ZP hardening, Fetub-/- derived oocytes cannot be fertilized by natural mating in vivo, nor by conventional IVF 8. Laser-assisted IVF, first perforates the ZP with the aid of a surgical laser, thus overcomes the zona as barrier and allows fetuin-B deficient oocytes to be fertilized. Blastocysts derived by laser-assisted IVF produced healthy offspring when transferred into wildtype foster mothers, demonstrating that ZP penetration is the critical step during fertilization in Fetub-/- mice. Fetuin-B may have a similar function in ZP hardening in humans. The fetuin-B gene is well conserved between mice and humans with 61% sequence homology (Fig. 4). Disulfide bridges are likewise conserved in mouse and human fetuin-B, suggesting a similar three-

14

Introduction dimensional protein structure. Domain D3 seems to be particularly exposed because antibodies from various sources rose against mouse or human fetuin-B had their binding epitopes mapped to domain 3 (Fig. 4) 34 suggesting that this domain is potentially an important part of its physiological function in reproductive biology.

Figure 4. Mouse and human fetuin-B alignment. Amino acid alignment of mouse fetuin-B (Q9QXC1) and human fetuin-B (Q9UGM5) shows a highly similar structure. Below the sequences, stars indicate amino acid identity, points mark amino acid distinction. Disulfide bridges are indicated by black lines connecting two cysteine residues (C, yellow highlighted) within one

domain or between the domains that are identified as D1 - D3. The leader peptide is gray highlighted, the fetuin motif LETDCHVL pink. Putative N-glycosylation sides are highlighted in orange; putative S-phosphorylation sides are written in bold. The antiserum K316 polyclonal rabbit anti-human fetuin-B binds weak (light blue) and strong (dark blue) to domain 3 of the human fetuin-B sequence. Commercial available antibodies against human fetuin-B from R&D

(Minneapolis, USA) bind to sequences in domain 2 and 3 (orange underlined). The antiserum K317 polyclonal rabbit anti-mouse fetuin-B binds also to sequences in domain 3 with different intensity; weak (light green) and strong (dark green).

1.2 Female (in)fertility Infertility is defined as the failure to become pregnant despite one year of regular unprotected intercourse with the goal of conceiving a child 35. Overall 10% to 15% of couples who wish to have children remain childless whereby males and females are equally affected 36,37. Assisted reproduction in humans started in 1978 when the first IVF child was delivered 38. The

15

Introduction

European Society for Human Reproduction and Embryology (ESHRE) estimated that the world's total number of babies born with aid of assisted reproductive techniques (ART) had reached 5 million in the year 2012 39. The most common ART are IVF and intracytoplasmatic sperm injection (ICSI). In general, in IVF the media contain oocytes and the capacitated (activated) sperm. Sperm attach and penetrate the ZP unaided. In contrast, in ICSI the single sperm reaches the oocyte by straight injection into the ooplasma. The number of ART treatments is ever increasing 40, partly because of lifestyle changes and career structures in industrialized countries require women to have children at older age. Overall fertility and natural conception rate both decrease with age, as does the rate in IVF and ICSI. In Europe, the pregnancy rate following IVF is 34.1% in women until 34 years-of-age and drops to of 26.1% in women at an age between 35 - 39. Women who are older than 40 years have mere 13.8% success rate 41. Comparable pregnancy rates were observed for women undergoing ICSI with 34.8% (≤ 34 years), 25.8% (35 - 39 years) and 13% (≥ 40 years) 42–44. From 1997 to 2004, it was noticed that the proportion of ICSI in comparison to IVF cycles, performed in infertility clinics, rose from 39.6% to 58.9% 45. The reasons were discussed controversially 45. It is argued that the number of ICSI cycles increased because of declining sperm quality; because men would like to have children also tend to be older. There is currently no scientific consensus when or why IVF or ICSI should be recommend. This probably leads to ICSI being the preferred method of choice in many fertility clinics, even if there is no strong indication of male infertility, just to be “on the safe side”. A mere 27.6% ICSI cycles were justified by a male factor alone, 38.4% by male and female mixed infertility and 11.7% by idiopathic infertility. Non-medical causes like the higher re-imbursement of ICSI over conventional ART may further enhance the preference of ICSI. This change in practice is viewed as critical by many because ICSI involves stronger technical interference with the natural fertilization process than conventional IVF does. Oocyte denudation, sperm immobilization and sperm injection are potential risks damaging the oocyte or embryo. For these reasons one should decide carefully which kind of artificial intervention is recommended.

1.2.1 Oogenesis The fundamental processes during oogenesis are similar in mouse and human. Oogenesis starts early in fetal development with the separation of germline cells from somatic tissue forming cells. The proliferation of gametes by mitosis proceeds and is completed during neonatal life. It is followed by a phase when the chromosomal recombination takes place and 16

Introduction the reduction from a diploid to a haploid genome occurs 46. Primary oocytes are growth- arrested in the first meiotic prophase (dictyate) until the puberty is attained. In sexually mature individuals, follicles regularly separate from the primordial follicle pool and differentiate into mature ovulatory follicles (Fig. 5).

Figure 5. Schematic overview of folliculogenesis. Primordial follicles are recruited from their pool and develop to primary follicles. The follicle size increases continuously during the maturation. At the stage of the secondary follicle the zona pellucida is assembled. Characteristically the preovulatory follicle forms an antrum containing the follicular fluid. The antrum increases in size until ovulation takes place and an oocyte with surrounding cumulus oophorus is released. Thereafter the resting cells in the ovary convert to the corpus luteum.

From primordial to primary follicles the size increases rapidly. This is caused by a high rate of mRNA and rRNA synthesis in the primary follicles fundamental for following oocyte development. During this secondary follicle phase the oocyte expresses, synthesizes and secretes ZP glycoproteins forming a glycoprotein matrix surrounding the oocyte 20,47. The ZP is formed between oocyte and granulosa cells. However, the mechanism of the ZP assembly at this stage is not well understood. Gap junctions between oocyte and granulosa cells and within the granulosa cells ensure the supply of the oocyte with amino acids and nucleotides.

17

Introduction

During the ongoing development, granulosa cells proliferate and adjacent to the membrana propia, surrounding the granulosa cell layers, theca cell layers are formed. When the follicular fluid appears within the granulosa cells the stage of the antral follicle is reached (Fig. 5). At this stage the oocyte is surrounded by a granulosa cell mass, named cumulus oophorus, located within the follicular antrum. The follicular fluid consists of secreted compounds of the granulosa cells and serum transudate 46. A close association of the number of granulosa cells and the follicular fluid estradiol concentration suggested that the main part of follicular fluid estradiol is granulosa cell derived 48. Additional studies demonstrated that a large portion of proteins in the follicular fluid is blood derived 49–52. Antral follicles with luteinizing hormone (LH) receptors at the granulosa and theca cells enter the preovulatory stage. Otherwise the antral follicle will become atretic and die. In the short time to ovulation the nuclear membrane breaks down and resumes the meiosis. The first meiotic division takes place and half of the chromosomes are located in the secondary oocyte while the other half is condensed in the first polar body. Immediately before ovulation the chromosomes in the secondary oocyte enter the second meiotic division and arrest during metaphase. The process of ovulation itself is triggered by the expansion of the follicular fluid pool and proteinases softening the follicle wall. Various proteinases including collagenases and gelatinases lead to a rupture and the oocyte, surrounded by cumulus cells, leaves the follicle ready for fertilization. The remaining follicle luteinizes to become a corpus luteum, the source of progesterone in the first embryonic days.

1.2.2 Hormonal regulation of oogenesis With sexual maturity, estrogen and gestagen secretion by the ovary increase. The endocrine activity of the ovary induces follicle development and thus the maturation to fertilizable oocytes. This process occurs periodically with two major phases. The first part, the follicular phase is primarily characterized by estradiol. The second part of the cycle after ovulation, the luteal phase, is dominated by progesterone (Fig. 6).

18

Introduction

Figure 6. Hormonal menstrual cycle regulation, schematic diagram. The menstrual cycle is divided into the follicular phase, characterized by a strong estradiol increase, and the luteal phase when progesterone dominates. Follicle stimulating hormone (FSH) is essential in the beginning of the follicular phase when follicles maturation starts while the luteinizing hormone (LH) peak is necessary to induce ovulation.

The pituitary gland secrets follicle stimulating hormone (FSH) and LH and thus induces the steroid production in the ovary. Mice lacking the pituitary gland showed follicle maturation but the number of follicles in different stages and the follicular DNA synthesis decrease emphasizing the role of gonadotropins (LH, FSH) in follicle proliferation and differentiation 53. At the beginning of each cycle LH binds at a low level to receptors of theca cells and thus induces the production of androgens. The granulosa cells require androgens, as well as FSH, to synthesize estrogen 54. Without the FSH increase the follicle development stops and follicles become atretic 55. Hence, FSH deficient female mice are infertile due to a blockade of follicular developmental before antral follicles are developed 56. During the follicular phase the increasing estrogen levels, in particular estradiol, lead to granulosa cell proliferation. A positive feedback mechanism ensures ever-rising estradiol levels (Fig. 6). In the clinic the serum hormonal change is commonly used in fertility control. Infertile patients who make use of assisted reproduction techniques like IVF or ICSI undergo controlled ovarian hormone hyperstimulation. Before hyperstimulation the ovarian function of patients has to be “down-regulated”. In general there are two different protocols using gonadotropin releasing hormone (GnRH)-analogues: the GnRH-agonists or the GnRH- antagonists. Repetitive GnRH-agonist application leads to a desensitization of the pituitary gland. In consequence the endogenous LH and FSH release is suppressed 46. In comparison GnRH-antagonist application inhibits the GnRH release directly by blocking the receptors 19

Introduction competitively. Consequently no endogenous gonadotropins are released. In general the controlled ovarian stimulation induces the synchronized development and maturation of a higher number of follicles. Usually recombinant FSH, urinary FSH or a combined medication of recombinant FSH and LH are used for hyperstimulation. Because of adverse reactions to hyperstimulation, regular medical checks are required throughout the treatment 57. Women undergo hormonal monitoring by blood sampling and an ultrasound scan of their follicles to monitor follicle size and maturation. Furthermore, the regular medical examinations are necessary because the hormonal treatments always constitute the risk of an overstimulation, which can range from mild side effects like nausea to serious secondary effects like the development and rupture of cysts. During usual ovarian hormone stimulation, appropriate estradiol concentration increase and follicle size should be attained 57. To select the optimum time for ovulation, both markers are monitored. Ovulation is ultimately triggered by human chorionic gonadotropin (hCG), a LH- analogue, which can be used to artificially induce ovulation. During the menstrual cycle elevated estrogen and FSH increase both LH (Fig. 6) and LH receptor expression in granulosa cells 46. Together this leads to much enhanced LH signaling, which triggers cytoplasmic and meiotic maturation of the oocyte. Further growth and endocrinology of the follicle cells affect an expansion of the follicular antrum along with loosening of the intercellular matrix of the granulosa cells. Simultaneously the granulosa cells switch from estrogen secretion to progesterone production. The progesterone synthesis is promoted by the increase of LH receptors. By a positive feedback of progesterone itself an increase can be observed (Fig. 6). The increasing progesterone level is important to suppress the maturation of other follicles and for ovulation itself. It was proved that the lack of progesterone or the use of progesterone antagonists inhibit ovulation 58,59. Hormone hyperstimulation before ART should never lead to spontaneous ovulation. To prevent a precocious ovulation, LH and progesterone are monitored 56. 36 hours after hCG injection follicles are aspirated under ultrasound guidance and then used for IVF or ICSI. If the fertilization is successful and the embryo implants in the uterus, pregnancy occurs and is associated with dramatic hormonal changes. Women produce more estradiol during pregnancy than women during the entire lifetime without being pregnant. Upon implantation the trophoblast (outer cells of the blastocyst) produces hCG that can be detected in the urine of the mother. Urine hCG which is diagnostic of pregnancy and hence the parameter tested in pregnancy tests 60. HCG is an LH-analog mediating synthesis of estradiol and progesterone in the corpus luteum. Hormone production prevents endometrium shedding and ultimately

20

Introduction abortion 61. With ongoing development of the embryo the placenta becomes the hormone- producing organ resulting in endometrium proliferation and development as well as an estradiol-induced increase in membrane potential, which protects the uterus against undesired contractions. The prevention of contractions are additionally supported by the increasing progesterone concentration 60. Moreover, the very high endogenous progesterone and estradiol concentrations prevent further ovulation during pregnancy. This situation is mimicked in contraceptive therapies. The daily intake of an estrogen and/or gestagen containing “pill” simulates a pregnancy precluding an “additional” pregnancy61.

1.3 Contraception Classical hormone-mediated contraception has intrinsic disadvantages despite proven success and worldwide application in human and animal medicine. In Germany more than 60% of the women who are currently using contraception take the contraceptive pill, in the United States on average every second woman 62. The contraceptive effect of hormonal birth control medication is based on the altered hormonal regulation caused by progestins (synthetic gestagens) and estrogens. The hormones influence the hypothalamus - pituitary gland - ovary interaction as well as the uterus and the tubes. Because they affect several important reproductive organs of the network simultaneously, they are reliable in pregnancy prevention. The commonly used hormone-mediated contraceptives are all based on the findings of Sturgis & Albright and Bickenbach & Paulikovics who described that repetitive administration of estradiol benzoate or progesterone inhibit ovulation 63,64. Oral formulations of estrogens and progestins down-regulate the endogenous secretion of both hormones. Simultaneously, the therapeutic estrogen and gestagen act synergistically in down-regulating gonadotropin releasing hormones by a negative feedback mechanism, thereby blocking LH and FSH secretion 57. Furthermore, due to the impaired LH and FSH secretion, follicle development and ovulation are inhibited. The low LH level additionally increases the viscosity of the cervical mucosa diminishing sperm motility and uterine tube motility, which is essential for oocyte transport. However, side effects are one of the main reasons for women to change the contraceptive strategy 62. Undesired effects of hormonal contraceptives include thrombosis, breast cancer, depression, bone fractures or a reduction in the ovarian reserve 65–69. Nevertheless, there is a urgent need for contraception to reduce the high number of unintended pregnancies noted in the US (50%) and worldwide (40%) 70,71. Contraception should prevent pregnancies, be 21

Introduction reversible as well as well tolerated. Apart from classical hormonal contraception targeting embryo implantation, recent contraceptive research attempted to prevent the fusion of sperm and oocyte. In female hamsters and mice monoclonal antibodies against ZP2 and ZP3 were administered to inhibit sperm binding 72–76. Alternatively ZP peptides were used for an active immunization 77. This treatment is, however, associated with the risk of autoimmune reaction and thus permanent infertility. Any novel contraceptive strategy must include a proof-of-principle study. The target has to be accessible for medication, the drug should bind target-specific and the biological role of the target should not replaced by other molecules 78. At best the target should be also tissue specific to avoid undesired target dependent side effects. In general, animal models are used to carry out toxicological studies and to validate a target before clinical studies start. For the development of human birth control medications these studies are only useful if the target has a human homolog.

1.4 Antisense oligonucleotides – chemistry and mechanism Antisense oligonucleotides (ASO) are short chemically modified oligonucleotides that bind their complementary target mRNA by Watson-crick base pairing. Usually they are 8 to 50 nucleotides in length (8- to 50-mer) 79. Oligonucleotides composed of a central block of phosphorothioate desoxyribonucleotides flanked by 2`-O-(2-methoxyethyl)ribonucleotides (2`-MOE) were called second-generation gapmer ASOs (Fig. 7). The centerpiece promotes RnaseH binding and thus degradation of the target mRNA (Fig. 8) 80 while the end sections protect the oligonucleotide from degradation 80,81. Phosphorothioate bonds between the nucleotides further increase the stability of ASOs 82.

Figure 7. 20-mer fetuin-B antisense oligonucleotide sequence, a second-generation antisense oligonucleotide. Chemically modified oligonucleotide (ION 637072) was produced by Ionis Pharmaceutical Inc., Carlsbad, USA. The characteristic gapmer structure is composed of ten central 2´deoxyribose nucleotides (squares) flanked by five 2´deoxyribose nucleotides (circle) on each side. Phosphorothioate bonds (red line) between the nucleotides as well as the gapmer structure and 5´methylcytosine modifications (purple shaded) increase the oligonucleotide stability. Regular nucleotides are shaded in blue.

22

Introduction

Said oligonucleotide modifications of second generation ASOs showed also a decreased immunogenicity compared to previous forms 83. The pharmacokinetic properties of phosphorothioate oligonucleotides have been well described. The bioavailability was up to 100% when oligonucleotides were injected subcutaneously 84. 2`-MOE gapmer bind by more than 90% to plasma proteins in mice and even higher in other species like monkey or 84. The oligonucleotides transfer from blood to the tissues and the distribution half-life was determined as one to two hours. After twelve hours, less than 1% circulate in blood while less than 5% are excreted indicating a good distribution into the tissues 85. Geary and colleagues showed that the same metabolites are found in the urine as in the tissue suggesting that the phosphorothioate nucleotides are degraded in the tissue by nuclease metabolism and as the same excreted 84. Clinical trials demonstrated efficacy of ASO therapy in rheumatoid arthritis, cardiovascular disease and diabetes 86–88. Due to their uniform chemistry, the pharmacokinetic properties of ASOs vary little across sequences and species 85,86,89–91. Systemically administered ASOs accumulate predominately in liver and kidney which in turn have shown most effected by off target effects 92. Both tissues are highly vascularized and have large numbers of tissue macrophages 92,93. Liver targeting is highly advantageous in the case of fetuin-B, because fetuin-B is a liver-derived plasma protein 3.

Figure 8. Antisense oligonucleotide mechanism. By complementary base pairing of the antisense oligonucleotide (ASO) to their target mRNA RnaseH binding is mediated. As indicated fetuin-B ASO (ION 637072) bind 70 bp at 5` side of the fetuin-B mRNA. The higher magnification depicts the fetuin-B ASO sequence (blue highlighted) and the matched fetuin-B mRNA sequence (black highlighted). Subsequent fetuin-B protein synthesis is inhibited by

mRNA degradation.

23

Introduction

1.5 The mouse as model organism in reproductive biology The laboratory mouse Mus musculus is one of the most common model organisms. In 2002 the first complete mouse genome was drafted 94. Since then, genomic data were compared with the human genome. Even if the mouse genome is slightly smaller at 2.5 gigabases (Gb) than the human genome (2.9 Gb), more than 90% of the murine genome sequence can be assigned into corresponding regions of the human genome and 80% of the genes are orthologous94. For this reason the mouse is often used for genetic ablation and mutation to investigate its functional role in diverse biological processes and diseases to transfer the knowledge to humans. In the field of reproductive biology over two hundred genetically modified mice with impaired fertility have been described 37,95. For example ZP2 as well as ZP3 deficient mice are infertile suggesting the essential role of the zona proteins in fertilization 96,97. Since ZP3 deficient (Zp3-/-) mice lacking a ZP, a genetic rescue strategy was chosen to study zona-sperm interactions in more detail. A humanized ZP with mouse ZP1, mouse ZP2 and human ZP3 was created 98. This model revealed that human sperm was unable to bind to the rescued humanized ZP while mouse sperm still did. Therefore, it was concluded that ZP3 is not involved in sperm binding. Moreover, the translation to human reproductive biology is an important aspect of humanized genetic mutations in mice. Basic research could be done in the mouse model without violating ethic principles. Avella and coworkers also used a mouse line with a humanized ZP (human ZP1, ZP3 and ZP4, lacking ZP2) and a previous established mouse line with all human zona proteins (ZP1 - 4) 99,100. By using the genetic mutated mice it was shown that neither human nor mouse sperm bound to the humanized ZP lacking mouse ZP2. These results proved that ZP2 is necessary for sperm binding in humans as well as species specificity of sperm binding to the ZP. In general, short reproductive cycles and robust reproduction render the mouse an attractive model organism for reproductive biology. The house mouse is a polyestrous mammal. One estrous cycle is with four to five days short compared to humans with 28 days. The four phases of the murine menstrual cycle can be distinguished by microscopic examining of the vaginal smear 101. The cellular composition of leukocytes, cornified and nucleated epithelial cells is characteristic for each phase. Proestrus and estrus are anabolic while metestrus is catabolic and diestrus a resting period of time 102. The mouse ovulates spontaneously, often around midnight. Different to humans, the corpus luteum is only build due to coital stimulation. Important reproductive data, representative for C57BL/6J mice, are summarized in table 1. Initially, this mouse strain was created by C. C. Little in 1921 103. Many substrains were derived which vary genetically. Beside C57BL/6J the most often used subline is

24

Introduction

C57BL/6N that was separated in 1951 104. These mouse strains differ in a mutation in the nicotinamide nucleotide transhydrogenase (Nnt) gene that is responsible for the glucose insensitivity and reduced insulin secretion 105. Overall the two inbred mouse strains are distinguished in 19 single nucleotide polymorphisms (SNPs) 106. There are several studies describing differences in behavior but not in respect to the reproductive behavior. Concluding reproductive data of C57BL/6J could be used exemplary for the mouse strain C57BL/6N 107– 109.

Table 1. Reproductive data of C57BL/6J mice. Star indicate data derived from ovarian hormone stimulated mice. Data is given as mean ± standard deviation 110–112,117.

Age at first litter, female 68.6 days

Time from mate to first litter 40.8 ± 10 days

Gestation duration 19.2 ± 0.3 days

Interval between litters 37.3 ± 6.2 days

Number of pups born in the first litter 6.9 ± 2.0 Number of pups born per litter 7.7 ± 1.8

IVF rate 66.2 ± 14%

Oocytes per donor* 25.2 ± 5.2

For female C57BL/6J mice the parental age at first litter was reported with 68.7 days 110. The gestation time is 19.2 ± 0.3 days 111 and the weaning age of the pups is reached with three weeks. On average it takes 38 days between the delivery of two litters demonstrating the laboratory mouse as a good breeder with short generation times 112. An important fact is that female mice show a postpartum estrous meaning that ovulation and the corpus luteum formation can take place already 12 - 18 hours after parturition 113. As a result lactation of the offspring and gestation of the next litter proceeds simultaneously. By hormonal regulation the gestation time can be prolonged, dependent on the litter size, from one to thirteen days 114. If mating occurs, female mice exhibit most times a vaginal plug that can serve as indication for mating. The plug, as firm whitish mass, is formed between vulva and cervix and stays usually for up to 24 hours, rarer up to 48 hours 115. It is composed by secreted proteins of the seminal vesicle and the anterior prostate of the male 116. Additionally, a very important aspect for female reproductive biology research is a high number of oocytes, which can be analyzed 25

Introduction simultaneously. By ovarian hormone stimulation on average 25.2 ± 5.2 oocytes per donor can be collected from one C57BL/6J mouse 117. Additional to the basic research in mice the mouse model is essential for pharmacological studies. The pharmacokinetic and pharmacodynamic effects of new medications and technologies can be analyzed to minimize the side effects and risks in humans. For example studies comparing the pharmacokinetic of ASOs in different species helped to develop effective medications to manage or heal diseases 84,86. Afore mentioned studies represent a critical step for the translation of new therapeutics to the clinical phase I studies in humans.

1.6 The aim of the study Recently, an essential role of fetuin-B in mouse fertility was demonstrated 8. The aim of this thesis was to confirm the role of fetuin-B in human reproductive biology. First, the fetuin-B regulation was investigated during the human menstrual cycle as well as in man. Next, it was hypothesized that fetuin-B could be a causative factor for at least some of the idiopathic infertile couples. Analysis of the serum fetuin-B in female infertile patients should provide information about an association between fetuin-B and fertility in women. The second aim of this thesis was to refine the knowledge of fetuin-B - ovastacin dependency in order to proof specificity of fetuin-B - ovastacin regulation of fertilization. This was studied in fetuin-B/ovastacin double deficient mice. Based on the positive outcome of this part, fetuin-B was investigated as a target for a non-hormonal contraception.

26

Experimental Procedures

2 Experimental Procedures

2.1 Fetuin-B in human reproductive biology

2.1.1 Ethical approval The study was approved by the local ethics committee of the RWTH Aachen University Clinics in accordance with the World Medical Association Helsinki Declaration on Ethical Principles for Medical Research Involving Human Subjects. Informed written consent was obtained from all participating volunteers and patients.

2.1.2 Study population and design In a longitudinally study serum fetuin-B was measured in five healthy male volunteers (34.3 ± 14.6 years). Furthermore, fetuin-B, luteinizing hormone (LH), estradiol (E2) and progesterone (P4) were determined in seven female volunteers with menstrual cycle (29.4 ± 4.1 years), and in four female volunteers on hormonal contraception (30.0 ± 6.5 years). The female volunteers took either no contraceptive drugs or mono or combined preparations containing 0 - 0.03 mg ethinyl estradiol, and 0.15 - 3.0 mg of various progestins as specified in the figure legend. Serum fetuin-B and sex hormones were also measured in one woman during and after pregnancy and in 25 patients undergoing ART with either gonadotropin stimulated IVF cycles (n = 21, 34.7 ± 4.2 years) or ICSI cycles (n = 9, 37.3 ± 4.8 years). Four patients had one IVF cycle followed by one ICSI cycle, respectively; one patient had two ICSI cycles. At least three blood samples of each IVF cycle were analyzed for the correlative analysis of serum fetuin-B, sex hormones and fertilization rate. IVF was successful (oocytes fertilized) in fifteen patients (34.4 ± 4.4 years), of which eight became pregnant; six patients (35.4 ± 3.3 years) failed IVF with no fertilized oocyte. Fetuin-B and albumin serum levels were also measured in patient-matched follicular fluids from nine patients undergoing oocyte aspiration (35.3 ± 4.0 years). Albumin content was taken as a proxy of overall protein concentration in the follicular fluid. The fetuin-B/albumin ratio was calculated in matched serum and follicular fluid samples to judge the volume distribution of these major plasma proteins. Pooled follicular fluid was obtained from five patients, and individual follicular fluids from altogether 69 follicles were obtained from 27

Experimental Procedures another five patients. Low grade cross-contamination of individual follicular fluid aspirates and contamination of the follicular fluid with small amounts of blood was controlled by routinely checking for the presence of hemoglobin, which would suggest blood contamination. Only samples with a hemoglobin content representing less than 250 erythrocytes/µl (Combur-Test, Roche Diagnostics, Mannheim, Germany) were used for the analysis.

2.1.3 Serum and follicular fluid sampling Blood was collected into S-Monovette serum tubes (Sarstedt, Nümbrecht, Germany) following venous puncture. After one hour clotting time, serum was separated by centrifugation (1500 x g, 10 min, 4°C). Serum was transferred to fresh microtubes, snap frozen in liquid nitrogen and stored at -20°C. The follicular fluid was aspirated during routine oocyte harvest by transvaginal ultrasound guided follicular puncture, and was cleared by centrifugation (200 x g, 5 min, 4°C). The clear supernatant was snap frozen in liquid nitrogen and stored at -20°C.

2.1.4 Protein and hormone measurements Fetuin-B was assayed in triplicates using a commercial sandwich enzyme-linked immunosorbent assay (ELISA) (human fetuin B DuoSet, R&D Systems, Minneapolis, USA), following the manufacturer's protocol. To ensure temperature consistency the microtiter plates (Nunc-Immuno MicroWell 96 well solid plates, Sigma-Aldrich, St. Louis, USA) were incubated between two adapted aluminum plates. E2, P4 and LH were measured by electrochemiluminescence immunoassay (ECLIA, Cobas Roche Diagnostics, Mannheim, Germany); human was measured using a colorimetric assay with bromocresol green (Cobas Roche Diagnostics).

2.1.5 Ovastacin inhibition assay The inhibition assay was performed by Konstantin Karmilin (Institute of Zoology, Cell and Matrix Biology, Johannes Gutenberg University Mainz, Germany; Director: Prof. Dr. Walter

Stöcker). The assay was used to identify the half maximal inhibitory concentration (IC50) of 28

Experimental Procedures ovastacin by recombinant human fetuin-B, (lot no. C050413-02, Invigate, Jena, Germany) and was done as previously described 8. In principle, fluorescence resonance energy transfer (FRET) was used to measure proteinase activity of ovastacin (540 nM) by a specific substrate

(Ac-RE(EDANS)-DRnLVGDDPY-K(DABCYL)-NH2, 25 µM). The fluorescence signal depended on the inhibitory capacity from fetuin-B to ovastacin. Human fetuin-B was tested in a range from 3.0 nM to 1.3 µM.

2.1.6 Statistical analysis Data were analyzed using GraphPad Prism 5.0c (GraphPad Software, San Diego, CA, USA) as detailed in the respective figure legends. Serum fetuin-B was analyzed by column statistics and values are given as mean ± standard deviation (SD). The Tukey method of plotting outliers was employed: outliers were defined as separated more than 1.5 interquartile distances IQR from the 75 percentiles. Coefficient of variation (CV) equals the standard deviation divided by the mean expressed as a percent. The Student´s t-test was used to compare fetuin-B levels in serum and follicular fluid. The Spearman correlation was used to study the correlation between fetuin-B and estradiol, progesterone and LH in untreated women and in women undergoing hormonal ovarian stimulation. Spearman correlation was also used to calculate the correlation between serum and follicular fluid fetuin-B. Serum fetuin-B levels are represented as measured with the exception of figure 15, where serum fetuin-B is represented as a linear regression of measured values. A p-value < 0.05 was regarded as statistical significant.

29

Experimental Procedures

2.2 Molecular fetuin-B probes - Monoclonal antibody production

2.2.1 Animal treatment All animal experiments were approved by the animal welfare committee of the Landesamt für Natur-, Umwelt-, und Verbraucherschutz (LANUV) of the state of North Rhine Westphalia. Animal maintenance, handling and treatment were performed according to the Federation for Laboratory Animal Science Associations (FELASA) recommendations. The mice were housed in controlled humidity and temperature on a constant twelve hours light dark cycle. Food and water were given ad libitum.

2.2.2 Production and purification of recombinant mouse fetuin-B The recombinant mouse fetuin-B was produced by Ralf Weiskirchen and Eddy van de Leur of the Institute of Molecular Pathochemistry, RWTH Aachen University as described previously 8,118. In brief, mouse fetuin-B cDNA was amplified, restriction sites introduced and afterwards cloned into pGEM-T vector (Promega, Madison, USA). Subsequently E.coli JM109 cells (Promega) were transfected for amplification. To introduce a histidine(His)-tag the fetuin-B fragment was transferred into a eukaryotic expression system (pcDNA3.1-V5-His, Thermo Fisher Scientific, Waltham, USA) and transposed into adenovirus (Ad5) by the use of an adeno shuttle vector (pΔE1sp1A). Further COS-7 cells were virus transduced. 72 hours before protein harvesting culture conditions were changed to serum free media to prevent contamination with bovine fetuin. The supernatant of COS-7 cells was 0.22 µm sterile filtered and purified using 5 ml HisTrap affinity columns (GE Healthcare, Little Chalfont, UK), packed with nickel-charged sepharose. The purification was done according to manufacturer´s protocol. All solutions and buffers were degased and 0.45 µm filtered. A constant flow rate of 5 ml/min was adjusted. To obtain the maximum yield imidazole (Applichem, Darmstadt, Germany) was added to the supernatant containing the recombinant protein in the same molarity as the binding puffer (20 mM sodium phosphate, 500 mM NaCl, 30 mM imidazole, pH 7.4). The same buffer with a higher imidazole concentration of 500 mM was used for protein elution. During the elution the flow rate was slowed to 1 ml/min and individual 1 ml fractions were collected. The protein purity was assayed by Coomassie blue staining and the presence of fetuin-B determined by Western blot. For Protein separation on a polyacrylamide gel samples were diluted in sample buffer (0.125 M TRIS/HCl, pH 6.8, 5% SDS, 10% glycerol, 10% 2- 30

Experimental Procedures mercaptoethanol, 0.01% bromophenol blue) and incubated for 5 minutes at 96°C. Finally 1.5 µl of each fraction was separated by SDS-PAGE (10% acrylamide, 5 x 8 x 0.1 cm3, BioRad, Hercules, USA) 119. The protein transfer onto nitrocellulose membrane (Amersham Protran, Sigma-Aldrich) was done using semi-dry electroblotting (Owl HEP-1, Thermo Fisher Scientific) at 1.5 mA/cm2 for 60 minutes. The membrane was blocked with 5% nonfat dried milk powder in PBS-T (Dulbecco’s phosphate buffered saline w/o Mg2+, Ca2+ supplemented with 0.05% Tween®20, Applichem) for 30 minutes at 37°C. Primary antibody (K319 polyclonal rabbit anti-mouse fetuin-B h3) was diluted 1:1000 in blocking solution and applied for 1 hour at 37°C. The second antibody (swine anti-rabbit IgG coupled with horseradish peroxidase (HRP), Dako, Glostrup, Denmark) was diluted 1:2000 in blocking solution and incubated at the same conditions. After both antibody incubations membranes were washed three times with PBS-T for 5 minutes. Bound antibody was detected by chemiluminescence in substrate solution (0.1 M TRIS/HCl, pH 8.5, 1.25 mM 3-aminopthalhydrazide, 0.45 mM p- coumaric acid, 0.015% hydrogen peroxide, all chemicals analytical grade by Applichem or Sigma-Aldrich) using a fluorescence imager (Fuji LAS Mini 4000). For protein purity evaluation the same samples as used for Western blot were applied on a polyacrylamide gel. Following protein separation gel was washed two times with distilled water and incubated for two hours in Coomassie staining (0.2% Coomassie brilliant blue, Serva, Heidelberg, Germany; 25% isopropanol, Applichem and 10% acetic acid, Applichem). Protein was visualized by washing the gel several times with 12.5% isopropanol and 10% acetic acid. For those fractions containing the purified protein a buffer exchange with TBS (10 mM TRIS, 20 mM NaCl, pH 7.5) was performed by gel filtration (5 ml ZebaTM spin desalting columns, Thermo Fisher Scientific). The protein concentration determination was done according to manufacturer´s protocol (PierceTM BCA Protein Assay, Thermo Fisher Scientific).

2.2.3 Immunization of fetuin-B deficient mice To produce monoclonal fetuin-B antibodies, four Fetub-/- mice were immunized with recombinant mouse fetuin-B. Pre-immune serum of these mice was sampled by puncturing the vena saphena immediately before the first injection was administered. Blood was clotted and centrifuged at 2000 x g for 10 minutes. Pre-immune serum was removed carefully and snap frozen using liquid nitrogen. For immunization the recombinant protein was mixed with

31

Experimental Procedures

TiterMax Classic (Thermo Fisher Scientific) until a homogeneous solution was obtained. Mice were injected intraperitoneally with a volume of 100 µl with 50 µg recombinant mouse fetuin-B, followed by four additional applications of 25 µg protein in approximately one- week intervals. After the third injection blood of each mouse was sampled again. Each serum was tested by an indirect ELISA as described in detail in 2.2.5 to validate the immunization success. Pre-immune serum of each mouse was used as negative control. With immunization cessation the three mice with the highest antigen detection signal, determined by ELISA, were transferred to BioGenes GmbH (Berlin, Germany) for hybridoma cell fusion.

2.2.4 Fusion of hybridoma cells The generation of anti-mouse fetuin-B immune globulin G (IgG) producing hybridoma cell lines was performed by BioGenes GmbH, Berlin, Germany. All used chemicals were analytical grade and were, unless otherwise noted, purchased from Merck, Darmstadt, Germany; Sigma-Aldrich or Riedel-de Haen, Seelze, Germany. The myeloma cell line SP2/0- Ag14 (SP2/0), purchased from Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ), was used for hybridoma cell production. SP2/0 were cultivated in DMEM (Dulbecco´s Modified Eagle Medium, PAA, Austria) supplemented with 10% fetal calf serum (FCS, PAA). Immunized mice were sacrificed by isofluran overdosing and cervical dislocation. Blood was sampled and serum was pooled serving as positive control in the following assays. The spleen was removed and homogenized. Spleen cells and myeloma cells SP2/0 were washed several times with DMEM and fused using polyethylene glycol (PEG 3350, 1 ml 50% w/v). Generated hybridoma cells were suspended in complete media with 20% FCS in HAT media (100 µM hypoxanthine, Gibco/BRL, 400 nM aminopterin, Sigma-Aldrich, 16 µM thymidine,

Gibco/BRL), seeded into eight 96-well plates (Corning-Star) and incubated at 37°C, 6% CO2, 96% humidity for ten days with two media changes. Peritoneal exudate cells were used as feeder cells.

32

Experimental Procedures

2.2.5 Specificity of hybridoma cell-derived monoclonal antibodies determined by ELISA The supernatant of the hybridoma cells was screened for fetuin-B specific IgGs using an indirect ELISA. Microtiter plates were coated with 50 µl/well of 2 µg/ml recombinant mouse fetuin-B (in 0.1 M carbonate-bicarbonate buffer, pH 9.6) and incubated over night at 4°C. Unbound antigen was rinsed off the plate by a washing procedure employing three washes with 380 µl TBS-T (TRIS buffered saline, 10 mM TRIS, 200 mM NaCl, pH 7.8, 0.01% Triton X-100). To avoid unspecific binding wells were blocked with 200 µl/well of TBS containing 2% bovine serum albumin (BSA, bovine albumin fraction V, Roth; Karlsruhe, Germany) for one hour at room temperature. Undiluted hybridoma cell supernatants were incubated with 100 µl/well for one hour at room temperature. Supernatant of SP2/0 cells served as a negative control, pooled antiserum of three immunized mice (1:100 in blocking buffer), harvested at the time of animal sacrifice and spleen removal, as positive control. Unbound primary antibody was rinsed of like unbound antigen, and secondary antibody (Fc-specific goat anti- mouse IgG coupled with HRP, diluted 1:5000, Pierce) was incubated for 1 hour at room temperature. Afterwards, unbound secondary antibody was rinsed by the washing procedure. Next, 0.1 mg/ml tetramethylbenzidine (TMB One Solution, Kem Tech, USA, in citrate/acetate buffer, pH 6.0 and 0.03% hydrogen peroxide, Applichem) was applied at 100 µl/well and color was developed for 20 minutes at room temperature. Reaction was stopped by 0.5 M H2SO4 at 50µl/well. The absorption was measured at 450 nm using a microplate reader. In addition, in house the produced antibodies were evaluated using the assay described above. To check antibody reactivity the supernatants were diluted 1:200 to 1:25,600. The Fc detection of the antibodies was done with a rabbit anti-mouse IgG/HRP (Dako). For preservation the supernatant containing the monoclonal antibodies was supplemented with 10 mM HEPES, pH 7.5 and 0.02% sodium acid.

2.2.6 Selection of stable antibody-producing cells After cell fusion, the hybridoma cells were cultivated for two weeks in selective medium. Aminopterin was then successively diluted-out by feeding the cultures with normal cell growth medium (DMEM). Hybridoma cells producing supernatant that scored a two-fold blank signal in ELISA were transferred into 48-well plates. Supernatant reactivity of these cells was retested by ELISA using a His-tag dummy protein as negative control. The 33

Experimental Procedures hybridoma selection time was minimized to prevent that nonspecific antibody secreting hybridoma cells would overgrow fetuin-B specific antibody-producing hybridoma cells.

2.2.7 Limiting dilution-cloning Hybridoma cells were cloned by two consecutive limiting dilution-cloning steps to separate antibody-producing cells from non-producing cells and to assure that the clones were monoclonal. After six to eight days cell proliferation was checked. Three days later the supernatants of proliferated cells were analyzed by ELISA. To determine the best subclone regarding proliferation and antigen detection the six best clones after each cloning were diluted into 24-well plates and tested multiple times to identify the best subclone to be used for the following cloning step and final cryopreservation. The heavy and light chains of the generated monoclonal antibodies were determined using a monoclonal antibody isotyping test kit (Serotec, Düsseldorf, Germany). All generated hybridoma cells were tested for mycoplasma contamination by Greiner Bio-One (Frickenhausen, Germany). Three hybridoma cells were generated and cultivated for bulk antibody production. Two hybridoma cells (20-4 and 27-10) were successfully bulk produced and purified using protein G sepharose purification. The purified monoclonal antibodies were stored in 200 mM glycine, 250 mM

NaCl, TRIS/HCl, pH 7.4, 0.02% NaN3. Hybridoma 15-1 did not generate enough antibodies to be bulk purified.

2.2.8 Cryopreservation of hybridoma cells During cloning, hybridoma cells were cultured in 24-well plates. For cryopreservation of back up cultures, cells from one well of a 24-well plate were harvested by centrifugation, and resuspended in 0.5 ml cryopreservation media (90% FCS, 10% dimethyl sulfoxide). Cells were immediately transferred into Nalgene® cryo-boxes with controlled freezing of 1°C each minute. After 24 hours cryo vials were stored at -80°C. After the second cloning step cells were cultured in 25 cm2 cell culture flasks. 1 – 3x106 hybridoma cells were cryopreserved in each aliquot, with four aliquots for each clone, similarly as described before.

34

Experimental Procedures

2.2.9 Specificity of hybridoma cell-derived monoclonal antibodies determined by Western blot Monoclonal antibodies were tested under the terms of reducing as well as under native conditions. Under reducing conditions the tertiary protein structure is denatured and antibodies can bind to continuous epitopes. Native conditions are used when the antibody recognizes discontinuous epitopes, where only in the folded protein the amino acid sequences are in close proximity to each other. Two different concentrations of recombinant mouse fetuin-B (0.3 and 0.6 µg) of different suppliers (manufactured as described in 2.2.2 and by Invigate, lot no. C110113) were applied to a polyacrylamide gel. Protein transfer and electroblotting were done as described in 2.2.2. The membrane was blocked with bovine serum albumin (1% BSA in PBS-T) for 30 minutes at 37°C. K319 polyclonal rabbit anti-mouse fetuin-B was used as positive control antibody, diluted 1:1000 in blocking solution and applied for 1 hour at 37°C. Supernatants of hybridoma cells or purified antibodies were diluted 1:200 and incubated at the same conditions. The secondary antibody (swine anti-rabbit or anti-mouse IgG coupled with HRP, Dako) was diluted 1:2000 (K319) or 1:500 (hybridoma cell supernatant) in blocking solution and incubated for 1 hour at 37°C. Washing steps and detection of the chemiluminescence signal were performed as mentioned in 2.2.2.

2.2.10 Immunoprecipitation of serum fetuin-B by monoclonal antibodies To evaluate the antigen binding capacity of the monoclonal antibodies an immunoprecipitation assay was performed. A checkerboard titration of both monoclonal antibodies mAb 20-4 and mAb 27-10 was used to find the optimal antibody ratio for mouse serum fetuin-B neutralization. It is likely that both antibodies bind to different epitopes. Therefore mAb 20-4 and mAb 27-10 were combined to increase the rate of antibody - antigen reactivity. The monoclonal antibodies were serially diluted in three steps and combined in all variations thus nine precipitation approaches were tested simultaneously. At first protein G sepharose (4 fast flow, GE Healthcare) was washed with PBS to remove the storage buffer. To immobilize the monoclonal antibodies 100 µl protein G Sepharose was incubated with 10 µl each mAb 20-4 and mAb 27-10 of different concentrations. Nine approaches with in total 2.75 µg - 11 µg combined monoclonal antibodies were incubated on a orbital shaker for three hours at 650 rpm and 4°C (ThermoMixer, Eppendorf, Hamburg, Germany). Antibody quantities are based on the assumption that murine serum contains 35

Experimental Procedures approximately 0.15 mg/ml fetuin-B, thus 15 µl serum comprises 2.25 µg of fetuin-B. If the antibodies bind 1:1 to the antigen, a volume of at least 6.75 µg monoclonal antibodies is required for precipitation considering that immune globulins have about a three times higher molecular weight (150 kDa) than the fetuin-B protein (~50 kDa). To assure a complete blending the reaction tubes were inverted several times during the incubation. Afterwards samples were centrifuged at 800 x g for two minutes and washed with 80 µl PBS. For the antibody - antigen reaction 15 µl mouse serum and 80 µl PBS were added and incubated for three hours at 650 rpm and 4°C. Afterwards the supernatant was removed carefully and the pellet was washed three times with PBS. To analyze the binding capacity of native serum fetuin-B each pellet and supernatant were analyzed by fetuin-B Western blot as described in 2.2.2.

2.3 Molecular fetuin-B probes - Fetuin-B as target for contraception 2.3.1 Generation of fetuin-B/ovastacin double deficient mice Ovastacin deficient FVB mice (FVB/129S6-Astl(tm1Dean), Astl-/-), were kindly provided by Jurrien Dean (Laboratory of Cellular and Developmental Biology, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD 25) through Walter Stöcker (Institute of Zoology, Cell and Matrix Biology, Johannes Gutenberg University Mainz, Germany). Seven Astl-/- female mice were mated with fetuin-B deficient (Fetub-/-) C57BL/6N male mice (B6-Fetb(tm,wja)) to generate Fetub-/-, Astl-/- double deficient mice by conventional breeding. Fertility of female mice was verified by mating Fetub-/- or Astl-/- single deficient or Fetub-/-, Astl-/- double deficient females with wildtype or Fetub-/-, Astl-/- males, respectively. Mating success was scored by daily palpation of the mice.

2.3.2 Fetuin-B and ovastacin genotyping Mice were genotyped by tail biopsy at weaning. Each tail tip was digested in a volume of 500 µl 0.1 mg/ml proteinase K (Thermo Scientific) in buffer (100 mM TRIS/HCl, pH 8.5, 5 mM EDTA, 0.2% SDS, 200 mM NaCl) and incubated over night at 55°C on a rotating rack. DNA was precipitated by adding the same volume isopropyl followed by a washing step with 70% ethanol. Afterwards the pellet was dissolved in 150 µl TE-buffer (10 mM TRIS/HCl, pH 8.0, 1 mM EDTA) and incubated for ten minutes at 55°C. For PCR reaction, resuspended genomic

36

Experimental Procedures

DNA, DreamTaq Green DNA Polymerase (Thermo Scientific), dNTPS and the following primers were applied: Fetub: P2 5’-CAAGTTCTAATTCCATCAGAAGC-3’ (Fetub-/- sense) P4 5’-GTCAGCTTCCACCTGACTCT-3’ (Fetub+/+ sense) P5 5’-AGAGCAAAATCCCCTGGTCA-3’ (5-10-5) Astl: P1 5`-AGG CCT TGT CAC CAG GTA TG-3` (Astl+/+ sense) P2 5`-GGGAGGATTGGGAAGACAAT-3` (anti-sense) P3 5`-GGGAGGATTGGGAAGACAAT-3` (Astl-/- sense)

A detailed description of fetuin-B and ovastacin genotyping PCR components is given in table 2.

Table 2. Fetub and Astl genotyping PCR components. stock working

concentration concentration Taq buffer 10x 1x dNTPs 10 mM 0.2 mM betain 5 M 1 M each primer 100 pM 0.5 pM Taq polymerase 5 U/µl 0.02 U/µl ultrapure water up to 24.5 µl DNA template 1.0 µl

Amplification conditions are provided in table 3 (Thermocycler T3, Biometra, Göttingen, Germany).

Table 3. Fetub and Astl genotyping PCR conditions. Step 2 to 4 were repeated for 30 cycles. temperature time 1. 94°C 5 min 2. 94°C 30 sec 3. 57°C 30 sec 4. 72°C 90 sec 5. 72°C 10 min 6. 4°C ∝

PCR products were applied on a 1.5% agarose gel and visualized by 0.002% ethidium 37

Experimental Procedures bromide (Sigma-Aldrich) supplementation. Using ultraviolet light, fluorescence of separated products was detected by Image lab software (Molecular Image Gel Doc XR+, BioRad). For fetuin-B amplicons of 130 bp (Fetub-/-) and 430 bp (Fetub+/+) were expected, for ovastacin amplicons of 400 bp (Astl-/-) and 800 bp (Astl+/+).

2.3.3 In vitro fertilization Oocytes from wildtype (Fetub+/+) and fetuin-B hemizygous (Fetub+/-) females showed no differences in IVF rate in previous studies 8. Thus wildtype and hemizygous-derived oocytes were pooled as “wildtype” to reduce the number of experimental . Two independent IVF studies were performed. One study was used to assess the in vitro fertilization rate in Fetub-/-, Astl-/- females (FVB and C57BL/6N mixed genetic background, n = 7). Female Fetub+/+ (C57BL/6N, n = 6) served as control. In another IVF study the permissive serum fetuin-B concentration required for contraception was determined in female Fetub+/+ mice (C57BL/6N, n = 18). Prior IVF, females were treated with ASO as described in 2.3.5. Six mice were each treated with fetuin-B ASO or control oligonucleotide. Six females were inseminated without any treatment (untreated control). The IVF protocol was performed as follows: Females were ovarian stimulated with 5 IU pregnant mare serum gonadotropin (PMSG, Intervet, Unterschleißheim, Germany) and 48 hours later with 5 IU human chorionic gonadotropin (hCG, Intervet). Thirteen hours after hCG application mice were sacrificed by isofluran overdosing and cervical dislocation. Ampullae were removed and cumulus-oocyte-complexes (COCs) were transferred to 200 µl human tubal fluid (HTF) medium (Tab.4).

38

Experimental Procedures

Table 4. Human tubular fluid (HTF) medium composition.

chemicals working concentration NaCl 5.9 g/l KCL 350.0 mg/l

MgSO4 x 7 H2O 49.0 mg/l

KH2PO4 54.0 mg/l

CaCl2 x 2 H2O 755.0 mg/l

NaHCO3 2.1 g/l glucose 500.0 mg/l Na-lactate (60% sol.) 3.4 ml/l Na-pyruvate 37.0 mg/ penicillin G 75.0 mg/l streptomycin 50.0 mg/l BSA 4.0 g/ phenol red (0.5% sol.) 0.4 ml/l

Media drops were covered with mineral oil and equilibrated with 5% CO2 overnight at 37°C. Outside the incubator COCs were kept in petri dishes at 37°C on a heated plate to minimize temperature fluctuation. One hour before COC removal, cauda epididymis of a male was removed and sperm were expressed in 90 µl modified Krebs-Ringer bicarbonate medium (Toyoda Yokoyama Hoshi + methyl-β-cyclodextrin, TYH+MBCD, Tab.5) for capacitation.

39

Experimental Procedures

Table 5. Modified Krebs-Ringer bicarbonate medium composition (Toyoda Yokoyama Hoshi + methyl-β- cyclodextrin, TYH+MBCD).

chemicals working concentration NaCl 6.98 g/l KCL 356.0 mg/l

MgSO4 x 7 H2O 293.0 mg/l

KH2PO4 162.0 mg/l

CaCl2 x 2 H2O 251.0 mg/l

NaHCO3 2.1 g/l glucose 1.0 g/ Na-pyruvate 55.0 mg/l penicillin G 75.0 mg/l streptomycin 50.0 mg/l methyl-ß-cyclodextrin 983.0 mg/l polyvinyl alcohol 0.4 ml/l

Sperm (1 - 2.5 x 106 /ml) were added to COCs and incubated for 3.5 hours. Subsequently oocytes were washed four times in 150 µl HTF medium to remove non-adhering sperm. After 24 hours the fertilization rate was assayed by the presence of two-cell embryos.

2.3.4 Antisense oligonucleotides chemistry and synthesis A fetuin-B 5-10-5 2-methoxyethyl gapmer with phosphorothioate backbone (5`-TAC ATT TCA TTG TGT GTG TC-3`, ION 637072), was synthesized and purified by Ionis Pharmaceuticals, Inc. (Carlsbad, CA, USA) as previously described 120. The synthetic ASO had a molecular weight of 7166.20 Daltons and 5` hydrogen and 3` hydroxyl caps. Lyophilized bulk was dissolved in phosphate-buffered saline (Dulbecco’s PBS w/o Mg2+, Ca2+) at 5 mg/ml or 10 mg/ml and sterile filtered using 0.22 µm pore size filter cartridges (VWR, Radnor, USA). ASO concentration was verified by absorbance measurement at 260 nm (BioSpectrometer, Eppendorf). A oligonucleotide 2´-MOE ASO 5`-CCT TCC CTG AAG GTT CCT CC-3` (ION 141923), of the same chemical and mechanistic class as the fetuin-B compound but not complementary to any known gene sequence, was used for control injections prior IVF.

40

Experimental Procedures

2.3.5 Serum fetuin-B down-regulation by ASO treatment Fetuin-B ASO (5 or 10 mg/ml in PBS) were injected subcutaneously into 6 - 13 week old C57BL/6N mice at 50 mg/kg or 100 mg/kg body weight. The control mice were treated with PBS or a control oligonucleotide. A single bolus was injected every other day for 14 days and every third day thereafter, when not stated otherwise. To determine serum fetuin-B concentration, blood samples were drawn from the vena saphena before, during and after ASO treatment.

2.3.6 Serum fetuin-B determination Blood was clotted and centrifuged at 2000 x g for 10 minutes. Serum was removed carefully and snap frozen in liquid nitrogen. For Western blots 0.3 µl murine serum was separated by SDS-PAGE. Assay was done as described in 2.2.2. Serum fetuin-B was quantified using purified recombinant mouse fetuin-B protein as a standard (lot no. C110113). Using Image J software densitometry of the chemiluminescence signal of standard and samples was performed and the serum fetuin-B concentration was calculated 121.

2.3.7 Mating of fetuin-B ASO-treated females Twelve female C57BL/6N mice (Fetub+/- and Fetub+/+) were injected with 100 mg/kg fetuin- B ASO and ten females with PBS. Twenty days after the first injection one PBS, one fetuin-B ASO-treated female and one male were co-caged. The PBS-treated females were used as internal control to prove fertility of the male and to exclude that the handling influenced the reproductive behavior. Mating success and pregnancy of females were scored by daily plug check and palpation. For plug analysis plug fragments were taken from the inner side of the vagina and dissolved in 10 µl PBS for microscopic analysis.

2.3.8 Offspring of fetuin-B ASO-treated females Litter size was noted at birth and weaning age. Three week old pups were weighed and blood was sampled retro-orbital. Serum fetuin-B was determined by Western blot as described in

41

Experimental Procedures

2.2.2. Serum fetuin-B concentrations were normalized to serum fetuin-B of an untreated three week old wildtype pup. All pups were genotyped for fetuin-B. To validate fertility of these pups females (Fetub+/-, n = 2) were co-caged with one male. Pregnancy and litter size were recorded.

2.3.9 Fetub and Gapdh RNA quantification After three weeks of PBS or fetuin-B ASO treatment, six females each group were sacrificed, and the liver, one kidney and one ovary were removed to determine the fetuin-B gene expression level. Organs were harvested into RNA stabilization reagent (RNAlater, Qiagen, Hilden, Germany) and incubated for 24 hours at 4°C. Afterwards organs were snap frozen using liquid nitrogen and stored at -20°C until RNA was purified. RNA was extracted (GeneJet Purification Kit, Thermo Fisher Scientific), followed by cDNA synthesis (Maxima first strand cDNA synthesis kit for reverse transcription quantitative polymerase chain reaction (RT-qPCR), Thermo Fisher Scientific) and RT-qPCR (Maxima SYBR Green/Fluorescien qPCR Master Mix (2x), Thermo Fisher Scientific) as described in the respective protocol. The following primers were used for RNA quantification: Fetub forward 5`-CAACATCTAACAACGTCTAGCC-3` and reverse 5`- GTAAGCCACTCTGCCAAATC-3`. Glycerinaldehyd-3-phosphate dehydrogenase (Gapdh) forward 5`-AGATGGTGATGGGCTTCCC-3` and reverse 5`- GGCAAATTCAACGGCACAGT-3` primers were used as housekeeping gene control. Each RT-qPCR reaction was performed in duplicates. For RNA amplification 7300 Real-Time PCR System (Applied Biosystems, Foster City, USA) was used. Initial denaturation (10 min, 95°C) was followed by 40 cycles of alternately denaturation for 15 seconds at 95°C and hybridization and elongation for 1 minute at 60°C. Primer specificity was verified by characteristic dissociation curve analyzed by 7300 System SDS software. For quantification Fetub values were normalized to Gapdh values.

2.3.10 Serum chemistry Serum was prepared from blood of females, which were used for Fetub and Gapdh mRNA quantification (2.3.9). Aspartate (AST) and alanine aminotransferase (ALT) were measured using a Vitros 350 analyzer and reagents (Ortho Clinical Diagnostics, New Jersey, USA)

42

Experimental Procedures performed by the laboratory of animal facility, university clinic RWTH Aachen, DIN EN ISO 9001:2008 certified.

2.3.11 Statistical analysis Data were analyzed using GraphPad Prism 5.0c (GraphPad Software, San Diego, CA, USA) as detailed in the figure legends. The two-tailed Student´s t-test was performed to compare litter sizes of PBS and ASO-treated mice. Two-tailed Student´s t-test was also used to analyze the weight of PBS and ASO-treated mice and the fetuin-B mRNA in different organs from both treatment groups. The two-tailed Mann Whitney t-test was used to compare the litter sizes of double deficient Fetub-/-, Astl-/- and single deficient Astl-/- females as well as for comparison of ALT and AST activity in sera of PBS and ASO-treated mice. A p-value < 0.05 was regarded as statistically significant.

43

Results

3 Results

3.1 Fetuin-B in human reproductive biology

3.1.1 Serum fetuin-B in humans Fetuin-B was monitored in human serum over the course of one month. Figure 9 shows that the serum fetuin-B levels varied between 2 - 5 µg/ml with an overall coefficient of variation of 19.4% in women (mean ± SD, 3.6 ± 0.7 µg/ml) and 33.3% in men (3.6 ± 1.2 µg/ml) (Fig. 9A and B, respectively). One male subject, demonstrating fetuin-B levels around 6 µg/ml at all time points, was identified as an outlier. The remaining four male subjects had serum fetuin-B levels of 3.0 ± 0.4 µg/ml (coefficient of variation 13.3%). Intra-individual serum fetuin-B coefficient of variation over the course of one month ranged from 7.1% to 24.0% in women, and from 4.9% to 13.0% in men. Collectively the data indicate a more constant hepatic expression of fetuin-B in men than in women.

Figure 9. Serum fetuin-B level variation over the course of one month (A) in female menstrual cycles (n = 8, seven individuals) and (B) in males (n = 5). Female cycle day 1 corresponds to day 1 of menstruation. Each curve represents one individual. The black lines represent the mean of all individuals.

3.1.2 Fetuin-B expression is stimulated by ethinyl estradiol Figure 10 illustrates that serum fetuin-B was elevated in women on hormonal contraception.

In women on medication with combined 17α-ethinyl estradiol (EE2) and progestin (Fig. 10A- C) serum fetuin-B levels increased, and rapidly decreased during the treatment-free interval suggesting steroid-dependent hepatic gene expression of fetuin-B. This steroid-induced increase in serum fetuin-B was observed in all women, irrespective of the basal serum fetuin-

45

Results

B level. Maximum serum fetuin-B levels reached 15 µg/ml (Fig. 10C) and thus up to 4-fold normal serum levels. Progestin alone (Fig. 10D) had no stimulatory effect, but a treatment change to an EE2 containing pill in the same volunteer stimulated serum fetuin-B expression (Fig. 10E).

Figure 10. Serum fetuin-B in women on contraceptive medication. Women received combinations of 17α-ethinyl estradiol (EE2) and progestin, (A) 0.03 mg EE2 and 0.15 mg Levonorgestrel, (B) 0.03 mg EE2 and 2 mg Dienogest, (C) 0.03 mg EE2 and 3 mg Drospirenon. (D) The woman took 2 mg Dienogest alone, followed by (E) 0.03 mg EE2 and 2 mg Dienogest. Combined EE2 and progestin stimulated serum fetuin-B (fetuB) expression, and serum levels dropped during the treatment-free interval. Progestin alone did not stimulate fetuin-B serum expression suggesting that the up-regulated fetuin-B expression by a combination contraceptive drug is caused by EE2. Serum estradiol (E2), progesterone (P4) and luteinizing hormone (LH) followed the expected patterns given the respective medication.

No causal relation between progestin intake and serum fetuin-B was observed in women on contraceptive medication. However, a correlation of endogenous progesterone and fetuin-B was determined during the menstrual cycle (Fig. 11A). The association suggested increase of serum fetuin-B in the luteal phase that could not be detected by fetuin-B measurements alone

46

Results

(Fig. 9A). Contrary no association between fetuin-B and estradiol or LH was observed in untreated women (Fig. 11B, C)

Figure 11. Matched serum fetuin-B and serum progesterone, estradiol and LH in untreated women. Serum fetuin-B correlates with (A) serum progesterone (P = 0.0009; Spearman r = 0.3314) and not with (B) estradiol (P = 0.0585; Spearman r = 0.1918) or (C) LH (P = 0.5625; Spearman r = -0.0592) during the menstrual cycle. Matched values of 7 untreated individuals (n = 98) are depicted. The lines depict linear regression curves with r2 = 0.073; P = 0.007 (A), r2 = 0.024; P = 0.132 (B) and r2 = 0.014; P = 0.240 (C).

3.1.3 Fetuin-B levels in serum and follicular fluid are tightly associated It was previously shown that fetuin-B is predominately synthesized in the liver 3. To exert its activity on oocytes fetuin-B must therefore traverse the follicle wall. Copious amounts of follicular fluid can be harvested during follicle puncture and oocyte isolation in ART patients. Follicular fluid fetuin-B level was compared to serum fetuin-B. Figure 12A illustrates the results of such a comparison in nine patients (patient 21 had two follicle punctures at different times), demonstrating that serum fetuin-B and follicular fluid fetuin-B were tightly associated in two thirds of the tested individuals. The tight association of follicular fluid fetuin-B with serum fetuin-B supports the view that serum fetuin-B is the source of follicular fluid fetuin-B. Three out of nine patients had significantly lower follicular fluid fetuin-B than serum fetuin-B (patient 9, 21 and 22). Nevertheless, the correlation (Pearson r = 0.87; P = 0.0012; r2 = 0.75; n = 10, Fig. 12B) of serum and follicular fluid fetuin-B indicate that serum fetuin-B passes relatively freely into the follicular fluid. Therefore, serum fetuin-B can be taken as a proxy of follicular fluid fetuin-B. Both serum and follicular fluid fetuin-B levels did, however, vary considerably between individual patients undergoing hormone treatment for ART.

47

Results

Figure 12. Association of serum and follicular fluid fetuin-B. (A) The intra-individual serum (Ser) and follicular fluid (FF) fetuin-B level were tightly associated in 6 out of 9 patients. (A, left) The FF of patients 9, 14, 15, 21, 46 were pooled during oocytes aspiration, (A, right) the FF of patients 21.1, 22, 48, 49 and 104 were aspirated individually. The fetuin-B concentration of pooled and individual FF was measured, and the mean value was calculated. (B) Correlation of serum and follicular fluid fetuin-B (P = 0.0012; Pearson r = 0.87; n = 10). Line describes linear regression curve with r2 = 0.75, P = 0.001). (C) Fetuin-B concentration of individual FF (black squares) and sera (open squares). (D) Fetuin-B to albumin ratio. Note that outliers detected in (C) do not differ in fetuin-B/albumin ratios. Two-tailed Student´s t-test; **P < 0.01, ***P < 0.001.

Figure 12C shows that follicular fluid isolated from several individual follicles derived from one donor all contained comparable amounts of fetuin-B. Patient 48 had one outlier follicle that contained substantially lower follicular fluid fetuin-B. Figure 12D shows that this difference vanished when the ratio of fetuin-B/albumin was determined in each follicular fluid. These values clustered even closer than the fetuin-B concentrations, suggesting that variations in follicular fluid fetuin-B reflect overall protein content which is known to increase with follicle maturation 51. Thus, the outlier follicle from patient 48 was most likely an immature follicle. 48

Results

3.1.4 Fetuin-B expression is stimulated by high endogenous estradiol during ovarian stimulation and during pregnancy The synthetic high potency estrogen 17α-ethinyl estradiol stimulated serum fetuin-B (Fig. 10). It was asked whether high endogenous estradiol would likewise stimulate fetuin-B expression. Statistical analysis of serum estradiol and serum fetuin-B in women undergoing spontaneous menstrual cycling showed no correlation (Fig. 11B). In contrast, the correlation between serum estradiol and serum fetuin-B became statistically significant (n = 142, 25 individuals; P < 0.0001, r = 0.414, Fig. 13A) at very high endogenous estradiol, like the ones attained upon hormonal treatment before IVF or ICSI.

Figure 13. Association of serum fetuin-B and serum estradiol and LH in women undergoing hormonal ovarian stimulation. Each data point represents matched serum fetuin-B and (A) estradiol or (B) LH and (C) estradiol and LH concentration, respectively. Values were measured during treatment cycles of fertility clinic patients (25 individuals). Spearman correlation shows association between fetuin-B and estradiol (A, P < 0.0001; Spearman r = 0.414; n = 142) as well as between fetuin-B and LH (B, P = 0.0032; Spearman = -0.249; n = 139) and between estradiol and LH (C, P < 0.0001; Spearman r = -0.353; n = 139). The lines depict linear regression curve with r2 = 0.249, P < 0.0001 (A), r2 = 0.075, P = 0.001 (B) and r2 = 0.081, P = 0.001 (C). In B and C three matched values could not depicted because LH values were below the detection limit.

The data presented here comprise an extension of published work. 122 Upon statistical re- examination of the published data set, serum fetuin-B was reverse associated with LH (P = 0.0032, r = -0.249, Fig. 13B). In any case, the reverse association of LH with fetuin-B is most likely not causal, i.e. due to fetuin-B regulation by LH, but rather coincidal, i.e. due to concommitant up-regulation of fetuin-B and down-regulation of LH upon hormonal treatment. LH secretion during the controlled hormonal ovarian stimulation is suppressed by GnRH down-regulation before the treatment started. Also for estradiol and LH a correlation 49

Results was determined (P < 0.0001, Spearman r = - 0.353 Fig. 13C) most likely due to coincidence as for fetuin-B and LH.

An extreme but physiological increase of endogenous estradiol occurs during pregnancy. To validate if serum fetuin-B also increased under very high physiological estradiol concentrations serum fetuin-B and estradiol were analyzed during the cause of pregnancy. Blood of a pregnant woman was sampled from the tenth week of pregnancy in two-week intervals (Fig. 14).

Figure 14. Association of serum fetuin-B and endogenous estradiol during pregnancy. Serum estradiol (E2, green) and serum fetuin-B (fetuB, black) of one woman is depicted during and after the pregnancy. Values revealed a single measurement of estradiol and the mean value of a triplicate measurement of fetuin-B.

The longitudinal study covered 30 weeks of pregnancy and a single measurement 15 weeks after delivery when endogenous estradiol had returned to normal values. Between week 10 and 16 both fetuin-B and estradiol increased. Serum fetuin-B nearly doubled from 5 to 10 µg/ml, estradiol rose from 1,000 pg/ml to 4,000 pg/ml. From then on estradiol increased continuously and reached a peak at 40,000 pg/ml immediately before delivery, while serum fetuin-B reached a plateau at week 16 onwards. After birth when estradiol returned to a normal level at 6 pg/ml during menstrual cycle, fetuin-B also rebound to normal female serum fetuin-B values of 2 µg/ml (Fig. 14). Thus serum fetuin-B differed considerably during pregnancy to serum fetuin-B after pregnancy.

50

Results

3.1.5 Serum fetuin-B is associated with fertilization rate in IVF Knowing that the lack of fetuin-B leads to premature ZP hardening in mice and thus to blocked fertilization, it was asked whether a similar correlation also exists in human patients undergoing IVF. To this end serum fetuin-B in patients undergoing ovarian stimulation was correlated with the fertilization rate in IVF. Successful IVF was defined as at least one oocyte fertilized during the respective cycle. Successful IVF procedures were associated with increasing serum fetuin-B (n = 15, P < 0.0001, Fig. 15A) during the ovarian stimulation. In contrast, fertilization failure (unsuccessful IVF = no oocyte fertilized) showed on average no increase of serum fetuin-B (n = 6, P = 0.118, Fig. 15B).

Figure 15. Association of serum fetuin-B and fertilization rate in IVF. (A) Serum fetuin-B increased upon ovarian stimulation in successful (at least one oocyte fertilized) in vitro fertilization (IVF) cycles (n = 15, P < 0.0001), (B) but remained unchanged in unsuccessful (no oocyte fertilized) IVF cycles (n = 6, P = 0.118). Endogenous estradiol increased in both (C) successful and (D) unsuccessful IVF cycles. Gray lines represented individual IVF cycles. Black lines depict the linear regression curves of mean serum fetuin-B and estradiol (P = 0.0019, C and P = 0.0254 ,D) respectively.

51

Results

Thus increased serum fetuin-B may be predictive of fertilization rate in IVF. Endogenous estradiol was routinely checked to control ovarian response to hormonal ovarian stimulation and increased in both IVF with fertilized oocytes and fertilization failure, respectively (Fig. 15C and D). In this study the fertilization rate only of the IVF procedure was studied. Serum fetuin-B in IVF stimulation was not correlated with clinical pregnancy, because the work in mice showed that fetuin-B is essential for fertilization, but not for later stages of pregnancy.

3.1.6 Ovastacin is inhibited by human fetuin-B In a final experiment the inhibitory potential of human fetuin-B on ovastacin was tested in vitro. Konstantin Karmilin (Institute of Zoology, Cell and Matrix Biology, Johannes Gutenberg University Mainz, Germany; Director: Prof. Dr. Walter Stöcker) performed the assays. Ovastacin activity was tested without or with recombinant human fetuin-B in a concentration range of 3.0 nM -1.3 µM (Fig. 16).

Figure 16. Human fetuin-B inhibited ovastacin. Reaction time with fetuin-B (vi) divided by the reaction time without fetuin-B (v0) is plotted against the molarity. Each data point represents mean of duplicate measurements. IC50 of 134 nM was calculated using GraFit 4 (Erithacus Software, Wilmington House, UK).

Human fetuin-B inhibited ovastacin with an IC50 of 134 nM suggesting a similar inhibitory 8 capacity compared to mouse fetuin-B (IC50 = 75 nM ). These in vitro data support the view that human fetuin-B may have a role in female reproductive biology as it was shown for mouse fetuin-B. 52

Results

3.2 Molecular fetuin-B probes - Monoclonal antibody production

3.2.1 Purification of recombinant mouse fetuin-B The production and purification of recombinant mouse fetuin-B was required for functional experiments analyzing the biological role of fetuin-B in fertilization.

Figure 17. Purification of recombinant mouse fetuin-B. Following purification of the recombinant protein (A) Coomassie staining and (B) fetuin-B Western blot was performed to judge yield and purity. Representative Coomassie staining and Western blot show the supernatant before purification (SN raw), flow, dead volume of pump and column from elution start, as well as ten further fractions. Molecular weight (kDa) is indicated at the right

Comparison of the Coomassie staining (unspecific protein staining) and the fetuin-B Western blot (specific fetuin-B detection) confirmed purity of the preparations. With both methods the same bands were detected indicating that only recombinant mouse fetuin-B was eluted (Fig. 17). Because the antibody-antigen detection by chemiluminescence is more sensitive than the Coomassie staining, the immunoreactive bands on the Western blot appeared thicker. The diffuse signal in fraction 2 to 5 (50 – 66 kDa, Fig. 17B) could be explained by low grade proteolysis of the recombinant protein despite the fact that the temperature was maintained at 4°C during the entire purification procedure. Fractions with a clear Coomassie signal (fraction 2 - 5) were pooled, buffer was exchanged by gel filtration, and the protein concentration was determined. In this particular run, 8.4 mg recombinant mouse fetuin-B (5.2 ml, 1.62 mg/ml) was purified. A total of about 27 mg recombinant fetuin-B was purified out of 2300 ml COS- 7 cell culture supernatant corresponding to roughly 10 mg/l supernatant.

Next, the recombinant fetuin-B was used in mouse IVF studies testing if the addition of fetuin-B to the IVF media increased the fertilization rate 123. 53

Results

Furthermore, 11.54 mg recombinant mouse fetuin-B was used for protein structure determination. The sample was sent to Tibisay Guevara (Proteolysis Lab, Department of Structural Biology, Molecular Biology Institute of Barcelona, Barcelona, Spain; Director: Xavier Gomez-Rüth), who performed crystallization experiments. As already known from inhibitory assays fetuin-B is a potent inhibitor of nephrosin, another member of the astacin family. On this account the crystallization of a fetuin-B - nephrosin complex was verified. Therefore a co-crystallization strategy was devised, for which I provided recombinant fetuin- B protein, and the Stöcker laboratory (Institute of Zoology, Cell and Matrix Biology, Johannes Gutenberg University Mainz, Germany) provided carp nephrosin. Around 500 different approaches with variations in pH, salt concentration and temperature ran to find the optimal crystallization conditions. However, no crystals could be observed to date, most likely due to the exceptionally high solubility of the proteins (personal communication Tibisay Guevara). Nevertheless, further trials are required to test if crystals may form at higher protein concentrations.

3.2.2 Monoclonal fetuin-B antibody production and evaluation Recombinant full-length fetuin-B protein was used to immunize fetuin-B deficient mice with the aim to produce monoclonal antibodies against mouse fetuin-B in mice. This strategy was chosen to produce antibodies that could later be used to deplete fetuin-B in mice without provoking an immunogenic reaction. To validate the immunization success of Fetub-/- mice, serum was sampled after the third immunization bolus. The sera of the immunized mice were tested for reactivity in indirect ELISA using the original antigen, recombinant mouse fetuin-B immobilized on the plate, and labeled rabbit anti-mouse IgG antibodies for bound antibody detection. Thus, sera were only tested for IgG antibodies, but not for IgM or any other subclasses that may have been elicited during the immunization of the mice. Pre-immune serum was used as negative control. The serum of all immunized Fetub-/- mice showed a strong absorption signal indicating antibody production against mouse fetuin-B (Fig. 18).

54

Results

Figure 18. Evaluation of antigen binding potential in serum of immunized Fetub-/- mice. Sera of four mice (#1 - #4) were tested in serial dilutions to detect antibodies binding to recombinant mouse fetuin-B. Pre-immune sera were used as negative control and showed no reactivity (data not shown). Optical density (OD) at 450 nm was subtracted by the blank at 570 nm. Each data point represents mean ± SD of duplicate measurements.

As expected, each mouse serum at the lowest serum dilution (1:200) yielded the highest signal. With serial dilution the absorption signal decreased. The pre-immune sera showed no absorption at 450 nm verifying that antibodies against fetuin-B were produced within four weeks of immunization (data not shown). Serum from mouse #2 showed the highest absorption at 450 nm indicating that blood from this mouse contained the highest anti-mouse fetuin-B IgG concentration and/or the highest affinity against the antigen. Consequently, this mouse had the best potential for the production of anti-fetuin-B antibody-producing plasma cells. Mice #1, #2 and #4 were shipped to BioGenes (Berlin, Germany) to generate hybridoma cells by fusing spleen cells from these mice with mouse B-cell lymphoma cells. Monoclonal antibody-producing hybridoma cell clones were generated by BioGenes employing routine protocols. Hybridoma supernatants were tested against recombinant mouse fetuin-B and recombinant human apolipoprotein B carrying a His-tag (abcam, Cambridge, UK). His- dummy protein was used to exclude His-tag-specific antibodies in ELISA. Wells with optical densities two times higher than the blank values were defined as positive (Tab. 6). Values are given as communicated by BioGenes.

55

Results

Table 6. Screening of primary culture cells following the selection phase. Supernatant of three wells containing primary hybridoma cells (15, 20, 27) were tested against recombinant mouse fetuin-B (rmfetuB) and against a His-dummy protein. All primary cultures show cross-reactivity with rmfetuB but no cross-reactivity with the His-dummy protein. As positive control pooled antiserum of immunized mice #1, #2 and #4 (AS 1:100) was used, the supernatant of SP2/0 cells served as negative control. OD 450 nm values are given.

Primary culture rmfetuB His-dummy protein 15 3.035 0.000 20 3.503 0.000 27 1.541 0.000 AS 1:100 3.505 0.214 negative control 0.000 0.001

The three primary hybridoma culture supernatants had at least twice the blank OD readings. The hybridoma supernatants showed no cross-reactivity against the His-dummy protein indicating that the produced antibodies reacted to an epitope within the recombinant protein and not the His-tag. The pooled polyclonal antiserum had high absorption values tested against the recombinant protein verifying the results depicted in figure 18. The pooled polyclonal antiserum showed also a slight cross-reaction against the His-dummy protein suggesting that the polyclonal antiserum contained also antibodies recognizing the His-tag as antigen. On the basis of these results, the three primary hybridoma cultures were cloned by limited dilution, and back up cultures were stored at -80°C. After the first cloning, all three primary cultures generated subclones producing antibodies against fetuin-B as determined by ELISA (Tab. 7).

Table 7. Cross-reactivity testing following the first cloning step. The supernatant of the subclones (15-1, 20-4 and 27-10) were tested against recombinant mouse fetuin-B (rmfetuB) and against a His-dummy protein. All subclones show reactivity with rmfetuB and no cross-reactivity with the His-dummy protein. Pooled antiserum of immunized mice #1, #2 and #4 (AS 1:100) was used as positive control, the supernatant of SP2/0 cells served as negative control. OD 450 nm values are given.

Subclone rmfetuB His-dummy protein 15-1 3.502 0.000 20-4 3.502 0.004 27-10 2.275 0.000 AS 1:100 3.500 0.435 negative control 0.014 0.000

Each supernatant producing subclone showed at least twice the blank signal. OD readings were scored as positive by BioGenes. However, the subclones 15-1 and 27-10 were not stable. Clones were declared stable when twelve out of twelve tested subclones were tested positive. 56

Results

Subclone 15-1 and 27-10 had to be subcloned one and two more times, respectively. Following the additional subcloning steps, referred to as second cloning step for all subclones, the cross-reactivity was evaluated by ELISA again (Tab. 8).

Table 8. Cross-reactivity testing following the second cloning step. The supernatant of the clones (15-1-1, 20-4-2 and 27-10-1) were tested against recombinant mouse fetuin-B (rmfetuB) and against a His-dummy protein. All clones show high reactivity with rmfetuB and weak cross- reactivity with the His-dummy protein. Pooled antiserum of immunized mice #1, #2 and #4 (AS 1:100) was used as positive control, the supernatant of SP2/0 cells served as negative control. OD 450 nm values are given.

Clone rmfetuB His-dummy protein 15-1-2 3.505 0.097 20-4-2 3.505 0.101 27-10-1 3.505 0.084 AS 1:100 3.500 0.435 negative control 0.028 0.000

All clones showed produced high ELISA signals indicating robust production of antibodies recognizing mouse fetuin-B. The supernatants were tested negative for the presence of mycoplasma (data not shown) and immunoglobulin subclasses were determined. All clones

(15-1-2, 20-4-1 and 27-10-1) had IgG1 heavy chain and κ light chain isotypes. To further evaluate antibody reactivity the supernatants were tested by indirect ELISA and Western blot under native and reducing conditions in house. Data are summarized in table 9, raw data are given in figure 19 and 20.

Table 9. Evaluation of antigen detection by the clones 15-1, 20-4 and 27-10. The antigen detection by the supernatant of these clones was examined by indirect ELISA and Western blot under native and reducing conditions (- = no signal; + = weak; ++ = good).

Clone 15-1 20-4 27-10 ELISA + ++ + Western blot (native) + ++ - Western blot (reducing) - - -

The antigen detection by indirect ELISA showed that signal intensity varied between the various clones. Clone 20-4 showed the highest absorption at 450 nm, clone 15-1 and clone 27- 10 had considerably lower absorption values (Fig.19).

57

Results

Figure 19. Evaluation of antigen binding potential in supernatant from monoclonal antibody- producing hybridoma cells after the second cloning. Clone 20-4 showed the strongest signal intensity at 450 nm (subtracted by the blank at 570 nm), followed by 15-1 and 27-10. With serial dilution the signal intensity decreased. Supernatant of SP2/0 cells served as negative control (NC) and showed no signal. Each data point represents mean ± SD of duplicate measurements.

In comparison BioGenes determined no differences between the clones following the second cloning step (Tab. 8). Discrepancies of measured values could be explained by different supernatant dilutions used in the assay. In the Western blot analysis (native conditions) the polyclonal anti-mouse fetuin-B antiserum K319 was used as positive control showing an oversaturated chemiluminescence signal against recombinant mouse fetuin-B (Fig. 20, top). Besides antiserum K319 only clone 20-4 showed a clear detection signal against non-reduced fetuin-B, while clone 15-1 reacted considerably weaker and clone 27-10 did not react at all. None of the monoclonal antibodies reacted with reduced fetuin-B in Western blots. In contrast, antiserum K319 was used as positive control under reducing conditions, and revealed a strong reaction against mouse fetuin-B.

58

Results

Figure 20. Evaluation of the supernatant of hybridoma cells after the second cloning by native and reducing Western blot. Supernatant of the hybridoma cells 15-1, 20-4 and 27-10 (diluted 1:200) were tested for detection of recombinant mouse fetuin-B (rmfetuB) produced in house (20130730) and supplied by Invigate (lot no. C110113). RmfetuB was applied in two concentrations (0.6 and 0.3 µg). K319 antiserum (polyclonal rabbit anti-mouse fetuin-B, 1:1000) was used as positive control showing an oversaturated chemiluminescence signal. Under native conditions (top) the supernatant of clone 20-4 shows the strongest detection signal, clone 15-1 a weak signal and it is absent for 27-10. Using supernatant of 20-4, 15-1 and 27-10 under reducing Western blot conditions (bottom) fetuin-B could not be detected at expected height.

In spite of varying degrees of antigen detection between the clones an antibody test production was commissioned. Clone 15-1-2 had a low antibody production rate thus no antibodies could be eluted during purification. The monoclonal antibody production of clone 20-4-1 yielded 1.6 mg/ml (13 ml, lot no. PP-030214-002) and 24 ml of clone 27-10-1 0.3 with 0.3 mg/ml (lot no. PP030214-003). The purified antibodies were compared to the raw supernatant before purification using an ELISA. As expected, the purified antibodies showed a stronger antigen detection signal than the supernatant of the cells after the second cloning step (Fig. 21).

59

Results

Figure 21. Antigen detection by the supernatant from monoclonal antibody-producing hybridoma cells, before and after antibody purification. ELISA measurements show the optical density (OD) at 450 nm (subtracted by the blank at 570 nm) of raw supernatant (SN) and after purification (pAb). After purification the signal intensity was stronger for both clones 20-4 and 27- 10. With serial dilutions the signal intensity decreased. Each data point represents mean ± SD of duplicate measurements.

The differences in signal intensity before and after purification is attributed to the fact that the purification led to an enrichment of the antibodies and thus improved antigen detection. Using the monoclonal antibodies in a Western blot under native conditions a diffuse signal could be detected (Fig. 22).

Figure 22. Evaluation of monoclonal antibodies 20-4 and 27-10 by native Western blot. Purified monoclonal antibodies 20-4 and 27-10 (diluted 1:200 / 8.0 µg/ml mAb 20-4 and 1.5 µg/ml mAb 27-10) were tested for detection of recombinant mouse fetuin- B (rmfetuB) produced in house (lot no. 20130730) or supplied by Invigate (lot no. C110113). RmfetuB was applied in two concentrations (0.6 and 0.3 µg). K319 antiserum (polyclonal rabbit anti-mouse fetuin-B, 1:1000) was used as positive control. Blurred chemiluminescence signal indicate a poor protein separation by the polyacrylamide gel. Thus evaluation of monoclonal antibodies is assessed with caution.

60

Results

Although the observed signals were vague it can be assumed that the antibodies bound specific to mouse fetuin-B. The signal decreased according to applied fetuin-B concentration (0.3 and 0.6 µg) and the signal intensity detected with both monoclonal antibodies is similarly as under the terms of reducing conditions. However, under reducing condition the usage of the monoclonal antibodies revealed stringent bands compared to the more diffuse bands observed by using a polyclonal rabbit anti-mouse fetuin-B antibody (Fig. 23).

Figure 23. Evaluation of monoclonal antibodies 20-4 and 27-10 by reducing Western blot. Both purified monoclonal antibodies (diluted 1:200 / 8.0 µg/ml mAb 20-4 and 1.5 µg/ml mAb 27-10) were tested for detection of recombinant mouse fetuin-B (rmfetuB) produced in house (lot no. 20130730) or supplied by Invigate (lot no. C110113). RmfetuB was applied in two concentrations (0.6 and 0.3 µg). K319 antiserum (polyclonal rabbit anti-mouse fetuin-B, 1:1000) was used as positive control.

This result revealed clearly the difference between monoclonal and polyclonal antibodies. While polyclonal antisera are a mixture of antibodies recognizing several epitopes and thus also various forms of the same protein, monoclonal antibodies were produced by one plasma cell clone targeting a single epitope. The antibody-antigen reaction could also be verified using an additional commercial available recombinant mouse fetuin-B (lot no. C110113) demonstrating the specific protein recognition. The evaluation in different assays of both monoclonal antibodies 20-4 and 27-10 are summarized in table 10.

61

Results

Table 10. Evaluation of antigen detection by purified monoclonal antibodies 20-4 and 27-10. The antigen detection by the purified antibodies was evaluated by ELISA and Western blot under native and reducing conditions. Clone 15-1 did not produce sufficient antibody quantity thus antigen recognition could not be determined (n.d.). (+ = weak; ++ = good; +++ = very good). Evaluation by the native Western blot is assessed with caution, indicated by brackets. In this assay protein separation did not work as usual.

clone 15-1 20-4 27-10 ELISA n.d. +++ +++ Western blot (native) n.d. (+) (+) Western blot (reducing) n.d. ++ ++

As mentioned before clone 15-1 produced an insufficient amount of antibodies preventing further analysis. Confirming previous results the antibodies derived from both clones 20-4 and 27-10 detected recombinant mouse fetuin-B using the ELISA. The absorption maximum is higher for clone 20-4, however, it has to be considered that the antibody concentration is about five times in 20-4 than in 27-10. In a final evaluation it is noted that both monoclonal antibodies detected the antigen mouse fetuin-B in different immunological assays.

To analyze if the monoclonal antibodies can be used to detect the native serum protein fetuin- B a preliminary immunoprecipitation assay with mouse serum was performed. With both monoclonal antibodies (mAb) 20-4 and 27-10 a checkerboard titration was done (Fig. 24A).

62

Results

Figure 24. Immunoprecipitation of mouse fetuin-B by monoclonal antibodies. (A) Various antibody quantities of 20-4 and 27-10 were tested in an immunoprecipitation assay followed by (B) mouse fetuin-B Western blot analysis. The concentration of the monoclonal antibody (mAb) 20-4 decreased from row A to C, mAb 27-10 from column 1 to 3. All antibody concentrations were pre- incubated with the same amount of protein G sepharose (proteinG). Following antibody immobilization the same volume of mouse serum was added into the reaction tubes. Pellet and supernatant (SN) of each approach were tested for mouse fetuin-B. Because of space reasons C2 was left out on the polyacrylamide gel. Molecular weight is indicated at the right.

An immunoprecipitation experiment was performed employing a range of 13 µg/ml to 51 µg/ml of monoclonal antibodies (150 kDa) to match the serum concentration of fetuin-B (~50 kDa) which was estimated to be ~10 µg/ml. This immunoprecipitation assay showed that the monoclonal antibodies recognized and bound native serum fetuin-B. The antigen-antibody complexes were harvested by centrifugation. The highest antibody quantity of 8 µg mAb 20-4 (Fig. 24B, row A) precipitated the most fetuin-B detected by Western Blot. With decreasing mAb 20-4 concentration (row A to C) a trend of decreasing mouse fetuin-B was observed in the pellet even if the signal intensity in C3 (pellet) differed. Ponceau S staining of the blot determined similar total protein amounts in each lane. In general, the amount of fetuin-B detected in the supernatant inversely correlated with the fetuin-B detected in the pellet. The fetuin-B enrichment in the pellet is more equal in each row (e.g. A1 to A3) than in each column (e.g. A1 to C1) indicating an association of immunoprecipitation with the 20-4 mAb concentration (Fig. 24B). Here, it has to be considered that the mAb concentration of 20-4 in reaction tube A1 is with 8 µg more than two times higher than mAb 27-10 with 3 µg. The monoclonal antibodies were originally meant for immunodepletion of fetuin-B in vivo. Simultaneous attempts to down-regulate serum fetuin-B in mice by ASO treatment proved highly successful. Therefore, the use of monoclonal antibodies to deplete endogenous fetuin- B in live mice was deferred.

63

Results

3.3 Molecular fetuin-B probes – Fetuin-B as target for contraception 3.3.1 Confirming fetuin-B as a target for contraception in vivo - Fertility recovery of Fetub deficient mice in Fetub/Astl double deficient mice Fetuin-B deficient (Fetub-/-) females are infertile (Fig. 25A). Additional ovastacin deficiency restored fertility despite fetuin-B deficiency. Fetuin-B/ovastacin double deficient (Fetub-/-, Astl-/-) females produced offspring confirming ovastacin proteinase as the prime molecular target of fetuin-B inhibitor.

Figure 25. Fetuin-B/ovastacin double deficiency restores fertility of fetuin-B deficient female mice. (A) Fetuin-B deficient mice (Fetub-/-, Astl+/+) were infertile. Double deficient mice (Fetub-/-, Astl-/-) had litter sizes comparable to ovastacin single deficient mice (Fetub+/+, Astl-/-), independent of the male genotype. Number of matings and litter numbers are given below the X-axis. (B) PCR genotyping proved fetuin-B (Fetub) and ovastacin (Astl) double deficiency of male, female and their offspring; +/+ = wildtype, +/- = hemizygous, -/- = deficient, H2O = no template control. The size of the PCR product is indicated at the right. Two-tailed Mann Whitney t-test; n.s. = not significant.

Consequently, in the absence of the target proteinase, the lack of the regulating inhibitor is of no further consequence. Figure 25A shows that the average litter size of Astl-/- females mated with double deficient males (4.7 ± 1.5) or wildtype males (5.3 ± 0.6) were similar to the average litter sizes of Fetub-/-, Astl-/- females mated with wildtype or double deficient males (5.6 ± 3.5 and 6.2 ± 3.7 pups per litter, respectively). Consequently the male genotype had no influence on the litter size. On average the litter sizes of Fetub-/-, Astl-/- and Astl-/- were also comparable to already published data of Astl-/- females 25,124. However, the variation in litter

64

Results sizes was much higher in female Fetub-/-, Astl-/- mice than in single deficient mice. PCR genotyping proved Fetub-/-, Astl-/- double deficiency of mated male, female and their offspring (Fig. 25B). The offspring appeared healthy at birth and attained similar weight at weaning age compared to the offspring of Astl-/- females (Fig. 26) indicating normal lactation of the females and a regular post-natal development.

Figure 26. Offspring of double deficient females had normal weight. Independent of the male and female genotype (Fetub-/-, Astl-/- or Astl-/-) the weight of their offspring was comparable. Mean litter size is indicated below the X-axis.

Based on the observation of a lower litter size in Astl-/- females following natural mating in this and previous studies 25,124, it was tested if Astl-/- females show a reduced fertility rate in vitro. IVF rates of Fetub-/-, Astl-/- mice were generally very high at 95 ± 1% compared with IVF rates attained with wildtype-derived oocytes at 61 ± 22% (Fig. 27A). The fertilization rate of wildtype-derived oocytes was comparable with published fertilization rates for C57BL/6J mice (66.2 ± 14%, see Tab. 1, page 23). Interestingly, following IVF of oocytes from Fetub-/-, Astl-/- females, unfertilized intact oocytes were never observed. If oocytes were unfertilized, they were damaged and contained coagulated cytoplasm and cell fragments. Reduced litter sizes after natural mating on the one hand and the exceedingly high IVF rates on the other hand indicated a defect in embryo implantation or development following successful in vivo fertilization. It is likely that the ZP of Fetub-/-, Astl-/- oocytes was the critical factor for an impaired fertility. Microscopic observation showed unusual high numbers of sperm attached to the ZP of Fetub-/-, Astl-/- oocytes (Fig. 27B). In addition, a high proportion of empty zonae (Fig. 27C, indicated by the arrow) and naked two-cell embryos was determined. In comparison, relatively few sperm bound the zona of wildtype oocytes 24 hours post insemination (Fig. 27D).

65

Results

Figure 27. In vitro fertilization rates of double deficient (DKO) and wildtype (WT) females. (A) Mean values of three independent experiments are depicted. Fertilization rate was determined by the presence of 2-cell embryos with or without zona pellucida (ZP- enclosed or ZP-free 2-cell embryos) 24 hours post insemination. Additionally the number of 1-cell oocytes and damaged oocytes were assayed. Exemplary micrographs of DKO (B, C) and WT-derived (D) oocytes are shown 24 hours post insemination. ZP-free 2-cell embryos and empty zonae (C), indicated by the arrow, could be detected for DKO oocytes, but not for WT oocytes. Scale bar indicate 100 µm.

3.3.2 Dose-finding study of fetuin-B ASO The recovery of fertility in the double deficient fetuin-B/ovastacin females underscored the decisive role of fetuin-B in fertilization, rendering fetuin-B a potential target for contraception. Therefore an attempt was made to regulate the fertility of female mice by 2`- MOE fetuin-B ASO-mediated fetuin-B down-regulation. Figure 28 illustrates an initial dose- finding study assessing the effect of fetuin-B ASO injection as single or repetitive boli. A single bolus of 50 mg/kg or 100 mg/kg body weight reduced serum fetuin-B to 90 ± 10% and 74 ± 3% of baseline level, respectively (Fig 28A). Control injections with PBS did not change serum fetuin-B. It took up to three weeks for the serum fetuin-B level to recover to baseline.

66

Results

Figure 28. Serum fetuin-B down-regulation by fetuin-B ASO. Serum fetuin-B measurements after (A) a single bolus or (B-D) repetitive boli of PBS, 50 mg/kg or 100 mg/kg body weight of fetuin-B ASO. Serum fetuin-B remained unchanged when only PBS was administered while it decreased to a minimum of 90 ± 10% (A, single bolus 50 mg/kg) or 74 ± 3% (A, single bolus 100 mg/kg). Repetitive boli of 50 mg/kg lowered serum fetuin-B to 49 ± 6%, 100 mg/kg boli to 33 ± 9% at day 15 (B). When administered at a higher frequency than depicted in (B), repetitive 50 mg/kg (C) or 100 mg/kg (D) boli decreased serum fetuin-B at day 20 to 33 ± 6 and 8 ± 6% of baseline levels, respectively. Depicted lines revealed the mean value with standard deviation of three (A, C), two (B), and ten mice (D) in each group. Arrows indicate time of bolus injection.

Repetitive injection of 50 or 100 mg/kg every other day and every third day from day 4 onwards, caused down-regulation of serum fetuin-B to 49 ± 6% and 33 ± 9% of baseline level (Fig. 28B). To further enhance the down-regulation the ASO boli were injected in intervals of two days until day 14. Ten repetitive boli of 50 mg/kg fetuin-B ASO reduced serum fetuin-B to 33 ± 6% of baseline level at day 20 (Fig. 28C). Despite serum fetuin-B down-regulation, the treatment with repetitive boli of 50 mg/kg fetuin-B ASO did not reliably prevent pregnancy. Six females were injected with ten repetitive 50 mg/kg ASO (n = 3) or PBS boli (n = 3) until day 20 (Fig. 28C). Thereafter, mating cages (n = 3) were assembled, each containing one PBS-injected and one ASO- 67

Results injected female mouse and one male. The mating behavior and pregnancies were scored by daily plug check and palpation.

Table 11. Mating behavior of fetuin-B ASO-treated females (repetitive 50 mg/kg boli). Three females in each group were treated with altogether ten boli of PBS or with 50 mg/kg fetuin-B ASO (ASO). At day 20 one female from each treatment group was co-caged with one male. Females were checked daily for vaginal plug and pregnancy. Numbers of vaginal plugs, pregnancy, gestation time and litter size were recorded to judge reproductive behavior.

Treatment Mating Vaginal Gestation Litter Pregnancy group cage plug (day) time size

PBS 1 21,25,27 1 18 6 2 21,24,37,50 0 0 0 3 21,32,36 1 19 9

ASO 1 21,23 1 20 3 2 25,38,52 0 0 0 3 23,28,29 1 19 2

All females, independent of PBS or ASO treatment, had repetitive vaginal plugs indicating regular mating (Tab. 11). Both females in mating cage 1 and 3 became pregnant and had litters within 25 days of cohabitation demonstrating fertilization success. The gestation time varied between 18 and 20 corroborating the published gestation time in C57BL/6J mice of about 19 days (see Tab. 1, page 23). No pregnancy could be observed for both females in mating cage 2 during 62 days of cohabitation suggesting an impaired fertility of the male. This preliminary study included three females per group and thus was relatively small. Obviously, ASO mediated down-regulation of serum fetuin-B to approximately 30% of baseline level was still sufficient to become pregnant. The litters of ASO-treated females with two and three pups tended to be smaller than the litters of the control females with six and nine pups (Tab. 11). Nevertheless, the females became pregnant indicating that the fetuin-B down-regulation was insufficient for contraception. To check if contraception could be reached by lowering serum fetuin-B concentration even further, fetuin-B ASO was injected as 100 mg/kg each boli. Ten repetitive boli of this increased dose down-regulated serum fetuin-B to 8 ± 6% of baseline level at day 20 day (Fig. 28D). An exemplary Western blot of the serum fetuin-B level during and after the ASO treatment with 100 mg/kg is shown in figure 29.

68

Results

Figure 29. Serum fetuin-B of PBS and ASO-treated females. Representative Western blot showing serum fetuin-B (fetuB) during (day 0 – 50) and after the treatment with PBS or fetuin-B ASO (ASO). During the treatment mice were injected with 100 mg/kg ASO or PBS every other day until day 14 and every third day thereafter. Blood sampling day is given above. In the PBS control group constant serum fetuin-B was determined while serum fetuin-B decreased when ASO boli were administered.

During the PBS treatment from day 0 to day 50, serum fetuin-B remained constant proving that simple injections did not influence the fetuin-B concentration. However, ASO treatment caused a continuous drop in serum fetuin-B during the first 20 days of treatment. Serum fetuin-B stayed low until the ASO treatment was completed at day 50. The recovery of the initial serum fetuin-B level took about 2 to 3 months suggesting a high in vivo stability of the ASO and/or a low fetuin-B expression rate.

3.3.3 Fetuin-B ASO mediated down-regulation causes infertility Figure 31A illustrates a typical serum fetuin-B down-regulation and recovery time course starting with 20 repetitive boli of 100 mg/kg within 50 days and the following washout period. Serum fetuin-B of all mice at day 0 was analyzed on a Western blot employing recombinant mouse fetuin-B as a standard (Fig. 30).

69

Results

Figure 30. Quantification of serum fetuin-B by recombinant mouse fetuin- B as standard. Representative Western blot shows serial dilution of recombinant mouse fetuin-B (rmfetuB, lot no. C110113, Invigate) and various serum fetuin-B samples of females before ASO treatment started (day 0). By densitometry measurements of standards serum fetuin-B was calculated. Letters above correspond to females depicted in figure 31. Molecular weight is indicated at the right.

Figures 31B - M show absolute serum fetuin-B concentrations determined by densitometry of twelve individual female mice undergoing ASO treatment and washout for the remaining time of the observation period of up to 226 days. Fertility of females treated with repetitive boli of 100 mg/kg fetuin-B ASO was verified by a mating study again.

70

Results

Figure 31. Fetuin-B ASO-mediated down-regulation of serum fetuin-B causes infertility. (A) Representative mouse fetuin-B Western blot showing serum fetuin-B (fetuB) level during (day 0 – 50) and after the antisense treatment. Blood sampling day is given above. (B – M) Serum fetuin-B concentration was determined individually for each mouse. During the treatment (continuous line) serum fetuin-B decreased in fetuin-B wildtype (K and L) and fetuin-B hemizygous (all others) mice. Mating (dashed line) was started at day 20. Black circle indicates vaginal plug but females became not pregnant. Gray circle indicates unstable plug. Vaginal plugs observed on consecutives days were plotted only once. Red circle symbolized the vaginal plug followed by pregnancy. All females had litters after treatment cessation, even if the plug when female conceived was not observed (J, L).

71

Results

After 20 days and altogether ten 100 mg/kg ASO boli, serum fetuin-B was down-regulated to 3 - 19 µg/ml. This corresponded to a 90% decrease compared to baseline. Fetuin-B down- regulation was consistent and comparable in all mice. Following treatment cessation it took 93 ± 19 days until serum fetuin-B returned to baseline. In contrast, PBS injection never down- regulated serum fetuin-B. PBS-treated females showed a constant serum-fetuin-B level during the entire treatment (Fig. 29, top). Upon mating, all PBS injected female mice became pregnant within thirteen days. This prompt fertilization success also proved that male mice were fully fertile, and that daily handling did not influence mating behavior. In contrast, 11 out of 12 fetuin-B ASO-treated females did not become pregnant during the treatment (Tab. 12). Figure 31M shows that this mouse had a comparatively modest serum fetuin-B level down-regulation to 19 µg/ml at day 20, while the remaining 11 mice had their serum fetuin-B level down-regulated to 7 ± 4 µg/ml.

Table 12. Mating behavior of PBS and ASO-treated mice (repetitive 100 mg/kg boli). Females were treated with PBS or fetuin-B ASO (ASO) and co-caged at day 20 onwards. Females were checked for vaginal plug, pregnancy and delivery. Litter sizes were recorded. Mating cage name corresponds to figure 31. Parentheses mark unstable plugs. For some females (PBS-treated in G and M) no vaginal plug was observed but females had litters. ASO-treated females in I and J were treated in a preliminary experiment and mated without PBS-treated female as control.

PBS-treated ASO-treated Mating Plugs in Litter Plugs in Litter Day of plug Day of plug cage total size total size B 21 1 6 43 (1) 0 C 21; 33 2 7 none detected 0 0 D 22; 23; 24; 30 3+(1) 5 29; 30 (2) 0 E 23; 24 1+(1) 7 none detected 0 0 F 24 1 3 none detected 0 0 G none detected 0 8 28; 29 (2) 0 H 24 1 9 21; 26; 27; 37 (4) 0 I - 28; 49; 50 3 0 J - 28; 29 2 0 K none detected 0 6 26; 40 2 0 L 21 1 9 none detected 0 0 M 23 1 7 36; 37 1+(1) 7

72

Results

Daily plug checks suggested that mating had regularly occurred. Unstable vaginal plugs were frequently observed in the ASO-treated females but not in PBS control mice (Tab. 12). Microscopic analysis of unstable plug material detected cornified epithelial cells suggesting that the mice were in estrus and ovulating (Fig. 32A). Sperm were also detected indicating mating despite unstable plug formation (Fig. 32B, C).

Figure 32. Microscopic analysis of an unstable vaginal plug during fetuin-B ASO treatment. (A-C) Micrographs show cornified epithelial cells, black arrows in B and C pointed at sperm present in the mucus.

A small mating study was performed to study if unstable vaginal plugs were associated with low serum fetuin-B. To this end the mating behavior, including plug formation and pregnancy, of Fetub+/- and Fetub-/- females were evaluated (Tab. 13).

Table 13. Mating behavior of fetuin-B hemizygous (+/-) and fetuin-B deficient (-/-) females. Females of both genotypes were co-caged with one male. By daily vaginal plug check and pulsation mating and pregnancy were noted. Vaginal plugs were observed regularly, independent of the genotype. Day on which females became pregnant is underlined. Fetub+/- females had litters, Fetub-/- were infertile.

Mating cage Fetub Complete Unstable Vaginal plug Litter/Size genotype plug plug (day) (number) (number) 1 +/- 1 1 3, 24 1/8 1 -/- 3 0 1, 18, 30 0 2 +/- 2 0 1, 21 2/11,7 2 -/- 2 0 15, 36 0

According to this study, serum fetuin-B was not associated with vaginal plug frequency or stability. As expected, Fetub+/- females had vaginal plugs, became pregnant and had litters, while all female Fetub-/- mice were infertile. Likewise Fetub-/- had vaginal plugs after mating,

73

Results despite them never becoming pregnant. Thus the observed differences in vaginal plug formation following fetuin-B ASO treatment were not caused by the reduced fetuin-B level, but likely were a side effect of the ASO treatment. Fetuin-B mRNA measured in liver tissue was tightly associated with the serum fetuin-B protein concentration strongly suggesting that reduced serum fetuin-B was caused by reduced hepatic synthesis (Fig. 33A). In ASO-treated animals (100 mg/kg body weight per bolus) fetuin-B liver mRNA dropped to 12.1 ± 3.1 % of values measured in PBS-treated mice. For good measure, fetuin-B mRNA was quantified in kidney and ovary. This was done to exclude major fetuin-B gene expression in these organs. Both PBS and ASO injected female mice expressed fetuin-B mRNA at very low level in ovaries (2.0 ± 1.2% vs. 2.3 ± 1%, n = 6 each group, Fig. 33A) and kidneys (0.6 ± 0.2% vs. 0.3 ± 0.1%, n = 6 each group) compared to liver fetuin-B mRNA expression in the same animals. These results suggest that fetuin-B mRNA expression in ovaries and kidneys is negligible compared to hepatic fetuin-B mRNA expression. Published literature reports ASO toxicity in the liver 91. The hepatotoxicity was assessed by measuring ALT and AST activity (Fig. 33B and C).

74

Results

Figure 33. Fetuin-B ASO led to reduced fetuin-B mRNA synthesis and increased aminotransferase activity. (A) Fetuin-B liver mRNA was significantly reduced in fetuin-B ASO-treated (ASO) females in comparison to PBS-treated females. Fetuin-B mRNA in both ovary and kidney is lower than in the liver of PBS-treated females, independent of the treatment group. Fetuin-B mRNA expression of 6 mice each group is normalized to fetuin-B liver mRNA of PBS- treated mice. Glycerinaldehyd-3- phosphate dehydrogenase (Gapdh) was used as housekeeping control gene. Two-tailed Student´s t-test; n.s. = not significant, **P < 0.01, ***P < 0.0001 (B) Alanine aminotransferase (ALT) and (C) aspartate aminotransferase (AST) activity was increased in females administered with fetuB ASO. Two-tailed Mann Whitney t-test; n.s. = not significant, **P < 0.01.

Figures 33B and C show that fetuin-B ASO treatment with repetitive 100 mg/kg boli increased ALT and AST activity in comparison to the PBS-treated mice. The elevated ALT and AST activities indicate modest liver toxicity, most likely caused by non-specific target side effects of ASO. Nevertheless, the females of both treatment groups appeared healthy, and were indistinguishable in behavior and weight (Fig. 34). The two-tailed Student`s t-test revealed no statistical differences in weight between both groups at any time point (day 0 - 23).

75

Results

Figure 34. Weight development of PBS and fetuin-B ASO- treated females. Females of both PBS and fetuin-B ASO (ASO) treatment groups had a comparable weight during the treatment. The weight was noted every two to three days. Each data point represents mean ± SD of 10 mice each. No statistical difference between both treatment groups was determined at any time point (two-tailed Student´s t-test).

Body weight of mice from both treatment groups was compared until day 23. After this date the PBS-treated females became pregnant. Therefore body weight was not any longer comparable with infertile ASO-treated females.

3.3.4 Contraceptive effect of fetuin-B ASO is reversible Next, it was tested whether the contraceptive effect of fetuin-B ASO injection was reversible. ASO treatment was terminated and mice were continuously mated. ASO-mediated down- regulation of serum fetuin-B and the associated infertility were completely reversible in that 60.3 ± 35.9 days after treatment cessation all mice had become pregnant (Tab. 14).

76

Results

Table 14. Reproductive behavior following ASO treatment cessation. Vaginal plug, gestation time and the number of pups at birth and weaning age are listed. Name of females correspond to figure 31. Female M became pregnant for the first time 13 days before treatment cessation. Vaginal plug in G was not detected but this female had also a second litter.

1st litter Day of 1st litter Day of 2nd litter 2nd litter at Gestation at Gestation Name plug (after (number of plug after (number weaning (d) weaning (d) treatment) pups) delivery of pups) age age

B 29; 30 19 5 4 1 28 9 9

C 43; 44 22 1 (dead) 0 2; 23; 24 19 7 7

57; 58; D 19 4 4 2 27 8 8 68;69

E 49; 50 20 4 4 1; 25; 26 19 9 9

F 22; 23 20 2 2 2 23 8 7

G 49; 50 19 4 4 - - 7 6

7 H 79; 134 22 7 +2(dead) 8; 9; 28; 4 K 29; 66; 19 4 1 30 5 5 +1(dead) 87; 103 88; 94; 95; L 105; 106; - 5 5 116; 117 M -13 20 7 7 1 30 8 6

Mean 20.0 4.3 3.8 25.1 7.6 7.1

Seven females had the second litter within 30 days after the first delivery, for three females the delivery of both litters were within 42 to 55 days, similar to published values for C57BL/6J mice of 37.3 ± 6.2 days (see Tab. 1, page 23). In two mice, copulation plugs matching the birth dates of litters were not observed, and are therefore also missing in figure 31J, L. Nevertheless, these females had obviously successfully mated.

77

Results

Figure 35. Fetuin-B down-regulation was reversible and did not affect the offspring. After fetuin-B ASO (fetuB ASO) treatment cessation all females had litters. (A) The first litters of ASO-treated mice were smaller than the litters of PBS-treated mice, the second litter sizes of ASO-treated mice were similar to the PBS group. (B) Body weight and (C) serum fetuin-B concentrations of three week old offspring were normal. Parent mice had been fetuin- B hemizygous, thus the offspring had all three possible genotypes. Black bars indicate fetuin- B wildtypes (Fetub+/+), gray bars fetuin-B hemizygous mice (Fetub+/-). Fetuin-B deficient pups had no serum fetuin-B. Serum fetuin-B was determined by normalization to an untreated 3 week old wildtype pup. Below the X-axis the number of pups and the respective mating cage are given. The mating cage corresponds to figure 31; for B, F and G two litters are shown. Two-tailed Student´s t-test; *P < 0.05, **P < 0.01.

Figure 35A shows that the first litter size of the fetuin-B ASO-treated females was smaller than the first litter of PBS injected mice (4.6 ± 2.3 vs. 6.7 ± 1.8 pups per litter, n = 20, P = 0.035). However, the smaller litter size was only temporary and the size of the second litter of fetuin-B ASO-treated mice was similar to PBS-treated mice (7.6 ± 1.3 vs. 6.7 ± 1.8 pups per litter, n = 18, P = 0.246). To reduce the number of experimental animals, PBS-treated females were only mated until they became pregnant for the first time. A normal gestation time with a mean duration of 20 days was observed for the first litter of fetuin-B ASO-treated mice (Tab. 14). Longer gestation times for the second litter were likely due to lactation 114. The weight of the first and second litter at weaning was 9.4 ± 1.4 g and 8.4 ± 1.4 g (Fig. 35B) and thus similar to values reported for the C57BL/6J strain by Jackson Laboratory (9.7 ± 1.9 g (male), 78

Results

9.3 ± 1.7 g (female)) 125. To study if serum fetuin-B was influenced by previous ASO treatment of the mothers, serum fetuin-B was measured in pups at weaning (3 week old). The parental animals were fetuin-B hemizygous and therefore the offspring had all three possible genotypes. Serum fetuin-B of the offspring corresponded to their respective genotype (Fig. 35C) in that no fetuin-B protein could be detected in serum from fetuin-B deficient pups, 68 ± 18% of wildtype serum fetuin-B was detected in hemizygous pups, and 95 ± 13% in fetuin-B wildtype pups. Mating of offspring from fetuin-B ASO-treated female mice proved that the offspring were fertile as well. In two mating cages Fetub+/- females had litters (7 and 8 pups) within five weeks of cohabitation.

3.3.5 Permissive serum fetuin-B range required for contraception The ASO mediated serum fetuin-B down-regulation was repeated with five mice, and oocytes were fertilized in vitro instead of natural mating. This experiment was designed to determine the permissive serum fetuin-B range required for contraception. After ovarian hormone stimulation cumulus-oocyte complexes were harvested and inseminated. Oocytes from six mice each that were treated with a control oligonucleotide or were left untreated were used as control. Both had unchanged serum fetuin-B levels (Fig. 36A). Fetuin-B ASO-treated females in this experiment attained serum fetuin-B of 9 to 18 µg/ml. The concentration range allowed studying the association of serum fetuin-B and IVF rate. The IVF rate was assessed by the presence of 2-cell embryos within 24 hours after insemination. Fetuin-B ASO-treated females had an IVF rate of 27 ± 18% (14% to 57% in individual mice, Fig. 36B). Thus the fertilization rate following fetuin-B ASO injection with 100 mg/kg body weight was generally higher in vitro than in vivo fertilization where only 1 out of 12 mice had become pregnant (8.3 %).

79

Results

Figure 36. Serum fetuin-B is associated with fertility. (A) Blood of female mice undergoing ASO treatment was sampled at the day of in vitro fertilization (females IVF) or during the ASO treatment (day 20, females mating). The number of fertilized oocytes/total number of oocytes is given in each bar representing serum fetuin-B of females used in IVF. Black bars indicate that fertilization occurred in vitro or in vivo, white bars indicate that no fertilization occurred. Letters below the X-axis correspond to figure 31. Serum fetuin-B (mean ± SD) of untreated hemizygous females (+/-, n = 16), control oligonucleotide (cOligo, n = 6) and PBS-treated females (n = 10) is depicted as control. (B) In vitro fertilization rate of oocytes from untreated, cOligo-treated and fetuB ASO-treated females. Each dot represents the fertilization rate of pooled oocytes of two females (untreated and cOligo) or of one female (fetuB ASO). Three independent experiments are depicted. (C) Fertilization rate of females with 0–10 µg/ml serum fetuin-B (n = 9), 11–20 µg/ml (n = 7) and 21–100 µg/ml (n = 6) are depicted. Data points represent individual fertilization rates; in the group of 21–100 µg/ml the fertilization rate of pooled oocytes (n = 2) is depicted. Bar at the right represents corresponding fertilization rate of each group in total.

To determine the minimum serum fetuin-B required for fertilization we compared the serum fetuin-B levels and matching fertilization rates of the in vivo and in vitro fertilization trials (Fig. 36C). In vivo mating without offspring was scored as 0% fertilization rate (Fig. 31B - L). The one in vivo mating that had produced offspring (7 pups, Fig. 31M) was scored as 100% fertilization rate. Combined values indicate that serum fetuin-B below 10 µg/ml should 80

Results be associated with infertility, serum fetuin-B between 10 to 20 µg/ml should be associated with highly variable fertility, and serum fetuin-B with 20 µg/ml and above should be associated with full fertility. This corroborates the previous observation during the in vivo mating study where all mice with serum fetuin-B of 10 µg/ml and below failed to become pregnant, and all mice with 20 µg/ml serum fetuin-B and over became pregnant (Fig. 36A) Collectively the data suggest that serum fetuin-B down-regulation below 10 µg/ml prevented pregnancy, and serum fetuin-B above 20 µg/ml permitted pregnancy. Oocytes from control oligonucleotide-treated mice had an IVF rate of 17 ± 8% (mean ± SD), which was lower that the IVF rate of oocytes from untreated mice at 40 ± 10% despite similar serum fetuin-B concentrations. While the IVF rate of untreated mice was in the expected range of fertile mice, the decreased IVF rate of control oligonucleotide-treated mice remains to be elucidated. It was noted that the percentage of fragmented and degenerated oocytes derived from control oligonucleotide-treated females was higher (41%) than in untreated (29%) or fetuin-B ASO-treated (3%) females. The AST and ALT activity in control oligonucleotide-treated mice were comparable to PBS-treated mice (Fig. 37) suggesting that the reduced fertilization rate in control oligonucleotide-treated females was not due to liver toxicity.

Figure 37. Alanine (ALT) and aspartate (AST) aminotransferase activity following PBS, control oligonucleotide or fetuin-B antisense treatment. ALT and AST were measured in serum after three (PBS) or four weeks (control oligonucleotide = cOligo). AST and ALT of fetuin-B antisense oligonucleotide (fetuB ASO) -treated females were comparable (ALT, P = 0.987; AST, P = 0.698) following three or four weeks of fetuB ASO treatment and are summarized. FetuB ASO- treated females showed significantly increased ALT and AST activity in comparison to cOligo and PBS-treated females. Each data point indicates single measurement of one mouse. Two-tailed Mann Whitney t-test; *P < 0.05, **P < 0.01.

81

82

Discussion

4 Discussion

4.1 Fetuin-B in human reproductive biology

Fetuin-B, a potent ovastacin inhibitor, prevents ZP hardening before fertilization and thus maintains oocytes fertilizable. The state of the zona pellucida is essential for IVF success both in humans and animals 23,126,127. ZP hardening occurs naturally following fertilization but also spontaneously during oocyte in vitro culture in mouse, rat, horse and human 28,126,128,129. Recently, it was described that premature ZP hardening also occurs in vivo in fetuin-B deficient mice leading to blocked fertilization 8. To determine the role of fetuin-B in human female reproductive biology, the fetuin-B level in serum and follicular fluid was measured. Fetuin-B levels in serum and follicular fluid were closely related suggesting that fetuin-B can freely diffuse from the blood into the follicles to act as an ovastacin inhibitor. This finding corroborates previous studies demonstrating that the overall protein concentration in serum and follicular fluid is comparable 50,51, and that most molecules up to 220 - 500 kDa can freely pass the blood-follicle barrier 49,130–132. Several studies reported that the composition of the follicular fluid indicates the maturation stage of the follicle 49,51. The association of serum and follicular fluid fetuin-B support this view, and renders serum fetuin-B a proxy of the fetuin-B concentration surrounding the oocyte. Male and female serum fetuin-B levels in healthy volunteers remained nearly constant over the course of one month. During the menstrual cycle serum fetuin-B, and LH or estradiol were not associated, while progesterone was. The association between fetuin-B and progesterone is most likely coincidental, not causal. This assumption is supported by the fact that serum fetuin-B was not increased when a healthy female volunteer took a synthetic progesterone alone. In contrast, the medication with ethinyl estradiol on hormonal contraception in healthy volunteers, and very high estrogen levels as shown for exceeding 600 pg/ml in ART patients undergoing hormonal stimulation and during pregnancy were associated with increased serum fetuin-B. However, one male showed serum fetuin-B up to 6 µg/ml, while the other males had 3 µg/ml but had no conspicuous changes in estradiol and LH arguing against exclusive estrogen regulation of fetuin-B expression. These data collectively suggest an indirect estrogen-mediated regulation of hepatic fetuin-B expression. An estrogen receptor binding site (GGTCANNNTGACC), which could potentially mediate the estrogen induction of fetuin- B, is present between the genetic loci of α2-Heremans-Schmid glycoprotein (AHSG) and

83

Discussion fetuin-B on human chromosome 3 133. However, this estrogen receptor-binding site is located ~8000 bp upstream of the fetuin-B gene transcription start site suggesting that major DNA conformational changes must occur for this site to exert a direct influence on the fetuin-B synthesis. Thus, a direct stimulation by estrogen of fetuin-B gene expression is unlikely. Instead very high estradiol concentrations, and the potent synthetic estrogen ethinyl estradiol may lead to signaling cross talk, and the activation of transcription factors other than the estrogen receptor. Studies in cultured human hepatoma cells indeed demonstrated that fetuin- B expression is induced by the farnesoid X receptor 134. Gene reporter studies involving estrogen and farnesoid X receptor agonists will might reveal if fetuin-B can be stimulated by this pathway, and if this can rescue female infertility. Furthermore, serum fetuin-B was increased during the hormone treatment of women undergoing successful IVFs, while it remained unchanged in patients with fertilization failure. Serum estradiol increased in both groups indicating an ovarian response to controlled ovarian stimulation. The endogenous estradiol concentration in patients with fertilization failure increased generally less than in IVF cycles with fertilized oocytes. However, the pilot study population was small and therefore results need further confirmation. The fertilization rate of 29% (6/21 IVF cycles) in these patients was well within range of published fertilization failure rates of 17 - 49% 135. Serum fetuin-B is a useful value in this respect because serum and follicular fluid are closely related suggesting that fetuin-B can freely diffuse from the blood into the follicles to act as an ovastacin inhibitor. Thus, fetuin-B behaved fundamentally different from its closest relative, fetuin-A, which was recently studied regarding its possible role in reproductive biology of women undergoing IVF 136. Similar to fetuin-B, fetuin-A concentration was high in follicular fluid of patients undergoing IVF. In contrast, fetuin-A was not associated with fertilization rate. This is in full agreement with the finding that fetuin- A deficient mice are fully fertile 5 and that fetuin-B, but not fetuin-A, in commercial ʹfetuinʹ preparations was responsible for the inhibition of ZP hardening 8,137–139. Upon ovarian stimulation of women undergoing IVF or ICSI, a reverse correlation between LH and fetuin-B was observed. It is most likely that the association is indirect by opposite regulation. Similarly, a reverse correlation between estradiol and LH was detected indicating a non-causal relation between LH and fetuin-B or estradiol. As during controlled ovarian stimulation LH is indepedent from changes in the estradiol concentration assured by desensisitation of the pituitary gland by GnRH agonists or antagonists before the stimulation started. Consequently upon hormone stimulation LH release is depressed despite increasing estradiol concentration. Nevertheless, at the very beginning of the hormone treatment LH is 84

Discussion still higher than during the stimulation and this behave opposed to the fetuin-B and estradiol level.

8 Recombinant mouse fetuin-B inhibited mouse ovastacin with an IC50 of 75 nM . A similar

IC50 was measured for human fetuin-B (134 nM). Given the molecular weight of fetuin-B (~50 kDa) and the respective serum concentrations in mice, fetuin-B is expected to act as potent ovastacin inhibitor in vivo as well. Adult mice had approximately 120 µg/ml serum fetuin-B, which equals 2400 nM. In comparison humans have 5 µg/ml serum fetuin-B equaling 100 nM. These numbers illustrate that the prevention of premature ZP hardening discovered in mice may be even more relevant to human reproductive biology, because serum fetuin-B concentration in humans in much closer to the IC50 than in mice. During the maturation process of human oocytes serum fetuin-B must be maintained at a level decidedly higher than the IC50 to ensure that follicular fluid fetuin-B keeps the oocyte fertilizable. It is hypothesized that slight changes in gene expression in fetuin-B will be functionally significant in humans, because the serum levels are close to the IC50 value and therefore fetuin-B might easily become a limiting factor.

85

Discussion

4.2 Molecular fetuin-B probes

4.2.1 Monoclonal antibody production The expression and purification of recombinant mouse fetuin-B were straightforward and reproducible. All required techniques and reagents are now established allowing high yields of recombinant protein. A functional assay using recombinant mouse fetuin-B during IVF studies showed that the supplementation of the media with fetuin-B improved the fertilization rate by inhibiting ZP hardening. Consequently, the recombinant protein can substitute the role of native fetuin-B during fertilization 123. The recombinant mouse fetuin-B was also used for the immunization of fetuin-B deficient mice. By generating hybridoma cells two types of monoclonal antibodies could be produced. During the production of the monoclonal antibodies the supernatant of the clones were evaluated by different labs and methods. The BioGenes company determined no differences between the clones following the second cloning step (Tab. 8). The reason is probably that BioGenes used undiluted supernatant to determine the antibody reactivity, while I used the supernatant at a dilution of at least 1:200. It is likely that the absorption at 450 nm of the undiluted hybridoma cell supernatant (BioGenes) was at a saturated level and thus differences in reactivity could not be detected. In general, clone 27-10 showed always the lowest antigen reactivity, which can be explained by a lower antibody concentration in the supernatant of the hybridoma cells indicated by the measurement of purified antibodies concentrations. The antibody production of clone 27-10 had to be assessed as less efficient. Consequently, much more supernatant had to be purified to obtain similar antibody amounts. Nevertheless, the preliminary immunoprecipitation results showed that a combination of both monoclonal antibodies 20-4 and 27-10 achieved partial neutralization of native mouse serum fetuin-B. For complete neutralization to occur, further titration experiments using varying amounts of both antibodies would be required. Because the concurrent attempts to down-regulate serum fetuin- B by ASO treatment proved highly successful, the use of monoclonal antibodies to deplete endogenous fetuin-B in vivo was deferred. This decision was taken in view of the very high antibody amounts that would be required for in vivo fetuin-B immunodepletion. Assuming that a mouse has a total blood volume of 3.0 ml and serum fetuin-B concentration is roughly about 0.15 mg/ml, a total amount of 0.45 mg fetuin-B would circulate in the body. Further assuming that a 1:1 antigen-antibody reaction takes place, for an antibody-mediated neutralization of serum fetuin-B to be successful, a single antibody bolus would have to comprise about 1.35 mg, because antibodies at 150 kDa have threefold higher molecular

86

Discussion weights than fetuin-B at ∼50 kDa. Furthermore, a single bolus would likely not be enough for complete in vivo immunodepletion, because hepatic protein synthesis would likely still proceed uninhibited. Therefore it was decided not to further pursue the fetuin-B immunodepletion approach, but to concentrate on ASO mediated down-regulation of fetuin-B instead. Since the liver is one of the organs with the highest ASO accumulation, fetuin-B as predominately liver-derived protein was a particularly promising target 92.

4.2.2 Fetuin-B as target for contraception The genetic ablation in the fetuin-B gene of female C57BL/6N, DBA/2 mice 8 as well as females with a mixed C57BL/6N, FVB background were infertile. Consequently, infertility of fetuin-B deficient females is background independent. Moreover, it was demonstrated that Fetub-/-, Astl-/- mice are fertile despite the absence of fetuin-B underscoring the specificity of fetuin-B as an inhibitor of ovastacin, and vice versa (Fig. 38).

Figure 38. Interaction of fetuin-B and ovastacin during fertilization. In wildtype mice (top) fetuin-B inhibits premature release of ovastacin thus preventing premature zona pellucida (ZP) hardening. In this situation sperm can fertilize the egg. Physiological ZP hardening occurs after fertilization and prevents further sperm penetration. In fetuin-B deficient mice (middle) precocious ovastacin release leads to premature ZP hardening. Sperm can not penetrate causing infertility. Fetuin- B/ovastacin double deficiency (bottom) exhibited a non-hardened ZP 8 and are fertile. (modified from )

87

Discussion

This in vivo gain of function experiment proved that premature ZP hardening did not occur in Fetub-/-, Astl-/- females. Furthermore, generation of double deficient mice provided genetic proof of a simple proteinase network consisting of a gelling substrate (ZP2) processed by a proteinase (ovastacin), which is regulated by a proteinase inhibitor (fetuin-B), collectively regulating fertilization. This proteinase network is reminiscent of certain aspects of blood coagulation/fibrinolysis where a gelling substrate fibrin(ogen) is created by the action of the proteinase thrombin(ogen) and thereafter removed by another proteinase, plasmin(ogen). Single deficiency of plasminogen was met with the expected thromboembolic disorder, while the loss of fibrinogen rescued mice from the pleiotropic effects of plasminogen deficiency 140,141.

It is shown that sperm can bind to the ZP of oocytes but not to 2-cell embryos 142. Several studies demonstrated that intact ZP2 was responsible for sperm binding, while ZP2 cleavage by ovastacin abrogated binding 25,99,100,143,144. Oocytes from Fetub-/-, Astl-/- mice bound higher number of sperm than oocytes from wildtype mice. This can be attributed to the fact that ZP2 stays intact in the absence of ovastacin, which in turn leads to persistent sperm binding even after fertilization. High numbers of 2-cell embryos in IVF of oocytes from Fetub-/-, Astl-/- mice suggested facile fertilization. Along these lines, Fetub-/-, Astl-/- female mice most likely have no defect in fertilization of their oocytes. The highly variable litter sizes following natural mating must therefore be due to losses of the early stage embryos. It is known that ZP hardening physically protects pre-implantation embryos during their passage through the oviduct to the uterus implantation site. In Fetub-/-, Astl-/- females this protection would be largely absent due to an unhardened ZP or the complete absence of a ZP suggested by the presence of zona-free 2-cell embryos 24 hours post insemination. Embryos with a non-hardened ZP are readily absorbed by the oviduct epithelium, a process also observed when the zona is mechanically removed or by disrupted ZP matrix in oocytes lacking ZP1 96,145,146. Recently, Winuthayanon and coworkers described impaired ZP2 cleavage in mice with conditional deletion of the estrogen receptor α (ERα). Oocytes with impaired ZP2 cleavage resulting in incomplete zona hardening were unable to develop into healthy embryos 147. Likewise, Gahlay et al. showed that Zp2Mut females had less blastocysts at embryonic day 3.5 despite similar numbers of 1- cell embryos compared to control females 144. An alternative explanation for the variable litter size in Fetub-/-, Astl-/- females would be

88

Discussion embryo abortion following polyspermy. However, my observations in Fetub-/-, Astl-/- oocytes suggest that they had a non-hardened ZP, but did not sustain polyspermy. Oocytes with mutated, non-cleavable ZP2 also had no supernumerary sperm in oocytes suggesting that post-fertilization polyspermy block was independent of ZP2 cleavage 144. Further work demonstrated that even in the complete absence of a ZP like in ZP2 (Zp2-/-) deficient oocytes, polyspermy was never observed 96. It has also been described that a post-fertilization block is executed at the oolemma, which prevents supernumerary sperm to fuse with the egg 148,149 . Thus the role of the ZP as a barrier to polyspermy is controversial, whereas the permissive role of ZP hardening for successful embryo development and implantation gains further support by this study. In summary, different mutated mouse lines proved the essential role of the ZP during folliculogenesis (Zp2-/-; Zp3-/-) 96,97, oocyte fertilization (Fetub-/-) 8 and passage down the oviduct (Zp1-/-; Zp2Mut; Fetub-/-, Astl-/-) 144,146. The absence of one or several of these players may impede proper ZP function leading into impaired fertility or complete infertility.

This study presents direct genetic and functional proof for the direct interaction of fetuin-B with ovastacin. Female Fetub-/-, Astl-/- were fertile despite the absence of fetuin-B, which leads to infertility in Fetub-/- single deficient mice. This important finding highlights fetuin-B as an interesting target for female fertility control. Confirming this hypothesis, fetuin-B ASO-mediated down-regulation of serum fetuin-B below 10 µg/ml in mice actually prevented pregnancy in vivo. The fact that fetuin-B is made in the liver, likely worked to the advantage of ASO-mediated down-regulation, because liver is one of the organs with the highest natural ASO accumulation, facilitating long-term efficacy of ASO treatment in vivo 92. With a single injection of 50 mg/kg fetuin-B ASO, a maximum down-regulation was achieved within three days. The depressed serum fetuin-B level returned to baseline after 20 days. Thus, the pharmacokinetic was comparable to other studies using ASO concentrations in a similar range 92. Furthermore, a direct association between serum fetuin-B and fertility was observed. In a preliminary mating study using ten repetitive 50 mg/kg fetuin-B ASO boli, females showed decreased serum fetuin-B with 33 ± 6 µg/ml, yet pregnancy occurred following natural mating. Likewise, pregnancy was observed in one female treated with ten repetitive 100 mg/kg fetuin-B ASO boli, leading to a down- regulation of serum fetuin-B to 19 µg/ml. In contrast, mice with serum fetuin-B lower than 10 µg/ml did not become pregnant upon natural mating. These observations suggested that serum fetuin-B of 20 µg/ml and more are sufficient for a successful fertilization, but depression

89

Discussion below 10 µg/ml are contraceptive. It is unclear, why one female mouse treated with repetitive boli of 100 mg/kg retained significantly higher serum fetuin-B level than 11 other mice treated identically. It was striking that after treatment cessation, serum fetuin-B increased considerably faster in that particular female mouse than in female mice that had been temporary infertile. Within one week the serum fetuin-B concentration had more than tripled. For comparison, females, which did not conceive had constant low serum fetuin-B levels during this time period. This observation suggests a diminished ASO uptake and/or accumulation, a faster reduction of ASOs or a higher fetuin-B expression in the one female that had become pregnant during ASO treatment. Independent of this outlier, fetuin-B down-regulation by the ASO technology was consistent and efficient. In all females a comparable pattern of fetuin-B down-regulation and washout period was observed. Serum fetuin-B was decreased to 10% of baseline level. For comparison, a recent publication studying in mice the role of fetuin-B in glucose metabolism, demonstrated that short hairpin RNA (shRNA) treatment decreased serum fetuin-B level to only 67% of baseline 150.

The variable IVF rate between 14 to 57% observed in oocytes derived from females having 9 - 18 µg/ml serum fetuin-B neatly covered the critical range of about 10 - 20 µg/ml serum fetuin-B, which seemed to be the tipping point for successful fertilization. Assuming a molecular weight of 50 kDa, 10 µg/ml serum fetuin-B corresponds to 200 nM. This value roughly corresponds to 2- to 3-fold the IC50 value of ovastacin inhibition by fetuin-B, which was determined at 75 nM in mice 8. In summary, at least in mouse oocytes, a serum fetuin-B concentration of 10 - 20 µg/ml may determine the success of in vivo and in vitro fertilization. This conclusion is further strengthened by a recent study demonstrating that IVF medium supplemented with 15 µg/ml recombinant mouse fetuin-B increased twice the fertilization rate of cumulus cell-free oocytes than fetuin-B free IVF medium 123. Premature ZP hardening is a common complication in assisted reproductive techniques. Therefore it might be worthwhile to use fetuin-B as a protein supplement in IVF media to prevent premature ZP hardening also in human IVF, and thus to increase the baby take home rate. An additional indication that serum fetuin-B concentration is associated with fertility is the reduced litter size in females treated with ten repetitive boli of 50 mg/kg fetuin-B prior to mating. Serum fetuin-B was reduced to approximately 30% of baseline. Nevertheless, females became pregnant, but importantly had only 3 and 2 pups in comparison to the PBS-treated females, which had litters of 6 and 9 pups and unchanged serum fetuin-B.

90

Discussion

Females used for IVF as well as the females during the mating study were repetitively treated with 100 mg/kg ASO boli. However, the mice whose oocytes were used in IVF had generally higher serum fetuin-B levels than the mice used of the mating study at day 20. One important difference between the two experimental groups of female mice was that mice destined for IVF were ovarian hyperstimulated while mice destined for natural mating were not. As mentioned above it is likely that the hormonal ovarian stimulation led to an increase of serum fetuin-B and thus ran counter to the ASO treatment. Besides the modest hepatotoxicity of ASO treatment observed in the mating study, an unexplained failure of IVF in oocytes derived from control oligonucleotide-treated mice was noted. ASO therapy is not generally believed to negatively influence reproduction. In previous studies the testis and the placenta showed only a low degree of ASO exposure 151,152. Only one case is reported where the application with second-generation ASO (TNFα inhibitor) led to a cell infiltration in the ovary associated with functional impairment of reproductive behavior by ovarian atrophy and a decrease in ovarian weight 153. However, TNFα itself plays a key role in regulating embryo development 154. Therefore it is unlikely that ASO side effects caused the impairment. The pathological changes were thus attributed to an increase in proinflammatory cytokines and chemokines rather than the toxic side effects caused by ASOs 155,156.

Published work investigated if ASO therapy of mothers would affect their offspring 151,152,157,158. The ASO dose of roughly 300 - 400 mg/kg/week in this study was high compared to the published studies (∼10 - 100 mg/kg/week). Therefore ASO toxicity in offspring of treated mothers was determined separately. Serum fetuin-B and body weight of pups at weaning were recorded. Body weight of the first and the second litter of fetuin-B ASO-treated mothers were comparable indicating normal lactation and post-natal development. Offspring of the first litter had a slightly higher mean weight. This was to be expected, because mothers took better care of smaller litters, which were typical of the first litter. Serum fetuin-B at weaning was also normal. Even if residual ASO passed through the placenta or reached the offspring by maternal milk it was too low to affect the serum fetuin-B of their pups. Nevertheless, a thorough investigation of the pharmacokinetic of fetuin-B ASOs was performed. After three weeks of ten repetitive 100 mg/kg fetuin-B ASO boli, efficient serum fetuin-B down-regulation was observed, but also a modest increase of the aminotransferases

91

Discussion

ALT and AST compared to the PBS or control oligonucleotide-treated mice suggesting a hepatotoxicity effect. The liver toxicity of ASO is sequence dependent explaining that two ASO sequences at the same concentration (e.g. fetuin-B and control oligonucleotide) may not have the same hepatotoxicity (personal communication Alfred Chappell). Even if the mice were indistinguishable from the PBS control mice in their behavior, a bolus of 100 mg/kg was high compared to other studies 91,152,159,160. Ongoing improvements in ASO technology may reduce the required dose and regimen. Recently, Prakash and coworkers showed that triantennary N-acetyl galactosamine (GN3)-tagged ASOs increased the down-regulating effect up to ten-fold in comparison to unmodified second generation gapmer ASOs 160. GN3- ASOs also did not increase transaminase levels. The higher efficiency is due to the GN3 modification, which greatly enhances liver parenchymal cell uptake through the asialoglycopotein receptor. For untagged ASOs, 80% of the ASOs were recovered from non- parenchymal cells, and only 20% in the protein expressing hepatocytes 159. An improved dose-response relation might reduce the unwanted side effects also in fetuin-B ASOs.

In summary, fetuin-B was established as a contraceptive target. The target is “druggable” by target specific ASOs. Within three weeks of ASO treatment a sufficient down-regulation could be achieved. Further it was shown that the down-regulation was reversible. After treatment cessation, serum fetuin-B returned to baseline serum levels, females became pregnant and had healthy and fertile litters. Fetub-/-, Astl-/- females were completely fertile. Thus ovastacin and fetuin-B are an unique functional proteinase/inhibitor combination playing an important role in fertilization. Obviously no other proteinase inhibitor can substitute fetuin-B deficiency in fertilization. Considering the high target specificity of ASO- mediated down-regulation, it is unlikely that hybridization of fetuin-B ASO with unintended targets, instead of the intended and observed fetuin-B down-regulation lead to the contraceptive effect of the ASO treatment 161. Consequently, important requirements of a new contraceptive target were met in this proof-of-principle study, which can thus serve as a basis for further research. Here, ASO-mediated down-regulation of serum fetuin-B and contraception were achieved in mice. Whether or not the contraceptive down-regulation of serum fetuin-B can be transferred to human biology needs to be confirmed. The fact that increasing serum fetuin-B during hormonal ovarian stimulation was associated with increased IVF rate in women raises hopes that the ovastacin/fetuin-B system may indeed also be exploited for contraception in humans. Interestingly, serum fetuin-B in humans is considerably lower (one tenth) than in mice 8

92

Discussion suggesting that fetuin-B ASO treatment might be more efficient in humans. In addition to fetuin-B ASOs, small molecule inhibitors of ovastacin may even be more viable alternatives for contraception. Small molecule proteinase inhibitors were recently introduced as blood clotting inhibitors 162,163. To sum up, it can be reasonably assumed that the fetuin-B/ovastacin interaction exists also in humans. The homology of mouse and human fetuin-B (61%) and ovastacin (69-78%) suggest that human fetuin-B is a potent ovastacin inhibitor in vivo, as it was shown for mouse fetuin- B 4,164,165. This hypothesis is supported by the first results of an in vitro inhibition from ovastacin by recombinant human fetuin-B. In line with this, the cleavage of ZP2 to ZP2f could be validated in human oocytes, which failed in IVF while immature oocytes possessed uncleaved ZP2 19. In conclusion, the human ovastacin/fetuin-B ZP2 proteinase network may serve as a druggable target for non-hormonal contraception in women.

93

94

Conclusion and Future Aspects

5 Conclusion and Future Aspects It was shown that fetuin-B plays a role in female reproductive biology in humans. Serum fetuin-B increased during successful IVF cycles, but did not change in unsuccessful IVF cycles despite increased estradiol levels. It is proposed that serum fetuin-B may be useful in predicting the fertilization success in IVF. Fetuin-B levels attained during ovarian hormone hyperstimulation may help to make an informed decision whether oocytes should be fertilized by IVF or by ICSI to overcome the ZP as a barrier. To validate this hypothesis individual follicular fluid fetuin-B need to be determined and correlated with the individual follicle fertility rate. In addition to the measurement of fetuin-B it would be interesting to learn more about its regulation. To this end fetuin-B promoter-reporter constructs must be cloned and studied in GFP-expressing reporter cell lines. The cells would be treated with different hormones and various concentrations to assess the induction potential of each factor. By fluorescence intensity direct information about the regulation of the promoter activity could be obtained. In regard to the fetuin-B regulation in humans it would be of interest why the up-regulation of fetuin-B during the controlled ovarian stimulation works better in some women than in others. Understanding fetuin-B gene regulation will further enhance our understanding of the biological role of fetuin-B in female reproductive biology. Structure/function analysis of fetuin-B will further enhance our knowledge of this novel target for fertility research. First, the structural elements should be defined that are necessary and sufficient for fetuin-B-mediated inhibition of ovastacin. To this end a series of domain- swapped or point-mutated versions of mouse and human fetuin-B should be tested in combination with ovastacin proteinase. Genetic analysis of human DNA should determine if any functionally relevant mutations exist in the of women, especially of women with idiopathic infertility.

My fetuin-B ASO mediated down-regulation and contraception study employed relatively high concentrations of ASOs. Next the amount of ASO and thus potential off target and side effects should be minimized using novel ASO derivatives with improved targeting and specificity including the recently developed GN3-coupled ASOs, whose hepatocyte-specific targeting was increased by an order of magnitude over conventional second generation ASOs.

95

Conclusion and Future Aspects

The generation of Fetub-/-, Astl-/- mice demonstrated that fetuin-B and ovastacin are unique interaction partners during fertilization. Following IVF, oocytes from Fetub-/-, Astl-/- females lost their ZP with relative ease. It should be determined if the loss of a protective ZP is also the reason that litter sizes varied greatly in naturally mated Fetub-/-, Astl-/- females, leaving ZP-free embryos without a protective shell and thus prone to premature hatching and resorption. The hardening status of Fetub-/-, Astl-/- derived oocytes should be determined by chymotrypsin digestion. IVFs in Fetub+/+ and Fetub-/-, Astl-/- should be performed, followed by oocyte in vitro culture until the blastocyst stage. If precocious hatching could be proved in vitro in Fetub-/-, Astl-/- oocytes, likely this process would also occur in vivo.

It is an ongoing discussion, whether in fact ZP hardening or the membrane block at the oolemma prevents polyspermy. Oocytes of Fetub-/-, Astl-/- females would be a great model for further investigations. In previous studies the ZP was removed mechanically or by a very low pH to distinguish between ZP and membrane block 166,167. Both treatments always entail the risk of oocyte damaging. The high number of 2-cell embryos 24 hours post insemination in Fetub-/-, Astl-/- derived oocytes suggested that polyspermy did not occur. Hence oocytes could be used to investigate the membrane block without ZP removal. In a first approach the number of sperm entry into the oocyte over time or by repetitive insemination has to be determined. This would prove if only the membrane block, independent of the ZP block, is essential to prevent polyspermy in mice.

96

References

References 1. Pedersen KO. Fetuin, a new globulin isolated from serum. Nature. 1944;154:575. 2. Lee C, Bongcam-Rudloff E, Sollner C, Jahnen-Dechent W, Claesson-Welsh L. Type 3 cystatins: fetuins, kininogen and histidine-rich glycoprotein. Front Biosci. 2009;14:2911-2922. 3. Denecke B, Gräber S, Schäfer C, Heiss A, Wöltje M, Jahnen-Dechent W. Tissue distribution and activity testing suggest a similar but not identical function of fetuin-B and fetuin-A. Biochem J. 2003;376:135-145. 4. Olivier E, Soury E, Ruminy P, et al. Fetuin-B, a second member of the fetuin family in mammals. Biochem J. 2000;350:589-597. 5. Jahnen-Dechent W, Schinke T, Trindl A, et al. Cloning and targeted deletion of the mouse fetuin gene. J Biol Chem. 1997;272:31496-31503. 6. Osawa M, Umetsu K, Sato M, et al. Structure of the gene encoding human alpha2-HS glycoprotein (AHSG). Gene. 1997;196:121-125. 7. Esemble genome browser. Transcript: FETUB-001 (ENST00000265029) - Homo sapiens. http://www.ensembl.org/Homo_sapiens/Transcript/Summary?db=core;g=ENSG00000 090512;r=3:186635969-186653141;t=ENST00000265029. Accessed February 11, 2016. 8. Dietzel E, Wessling J, Floehr J, et al. Fetuin-B, a liver-derived plasma protein is essential for fertilization. Dev Cell. 2013;25:106-112. 9. Stöcker W, Karmilin K, Hildebrandt A, et al. Mammalian gamete fusion depends on the inhibition of ovastacin by fetuin-B. Biol Chem. 2014;395:1195-1199. 10. Heiss A, DuChesne A, Denecke B, et al. Structural basis of calcification inhibition by alpha2-HS glycoprotein/fetuin-A: formation of collodial calciprotein particles. J Biol Chem. 2003;278:13333-13341. 11. Li W, Zhu S, Li J, et al. A hepatic protein, fetuin-A, occupies a protective role in lethal systemic inflammation. PLoS One. 2011;6:e16945. 12. Wang H, Li W, Zhu S, et al. Peripheral administration of fetuin-A attenuates early cerebral ischemic injury in . J Cereb blood flow Metab. 2010;30:493-504. 13. Schäfer C, Heiss A, Schwarz A, et al. The serum protein α2–Heremans-Schmid glycoprotein/fetuin-A is a systemically acting inhibitor of ectopic calcification. J Clin Invest. 2003;112:357-366. 97

References

14. Herrmann M, Schäfer C, Heiss A, et al. Clearance of fetuin-A-containing calciprotein particles is mediated by scavenger receptor-A. Circ Res. 2012;111:575-584. 15. Hogan B, Beddington R, Constantini F, Lacy E. Manipulating the Mouse Embryo. A Laboratory Manual. 2nd ed. New York: Cold Spring Harbor Laboratory Press, Cold Spring Harbor; 1994. 16. Wessling J. Gene Deletion and Functional Analysis of Fetuin-B. Dissertation. RWTH Aachen University. 2007. 17. Hughes DC, Barratt CLR. Identification of the true human orthologue of the mouse Zp1 gene: evidence for greater complexity in the mammalian zona pellucida? Biochim Biophys Acta. 1999;1447:303-306. 18. Lefièvre L, Conner SJ, Salpekar A, et al. Four zona pellucida glycoproteins are expressed in the human. Hum Reprod. 2004;19:1580-1586. 19. Bauskin A, Franken D, Eberspaecher U, Donner P. Characterization of human zona pellucida glycoproteins. Mol Hum Reprod. 1999;5:534-540. 20. Wassarman PM. Zona pellucida glycoproteins. Annu Rev Biochem. 1988;57:415-442. 21. Green DP. Three-dimensional structure of the zona pellucida. Rev Reprod. 1997;2:147- 156. 22. Greve JM, Wassarman PM. Mouse egg extracellular coat is a matrix of interconnected filaments possessing a structural repeat. J Mol Biol. 1985;181:253-264. 23. Braden AWH, Austin CR, David HA. The reaction of the zona pellucida to sperm penetration. Aust J Biol Sci. 1954;7:391-409. 24. Bleil JD, Beall CF, Wassarman PM. Mammalian sperm-egg interaction: fertilization of mouse eggs triggers modification of the major zona pellucida glycoprotein, ZP2. Dev Biol. 1981;86:189-197. 25. Burkart AD, Xiong B, Baibakov B, Jiménez-Movilla M, Dean J. Ovastacin, a cortical granule protease, cleaves ZP2 in the zona pellucida to prevent polyspermy. J Cell Biol. 2012;197:37-44. 26. Inoue M, Wolf DP. Solubility properties of the murine zona pellucida. Biol Reprod. 1974;10:512-518. 27. Drobnis EZ, Andrew JB, Katz DF. Biophysical properties of the zona pellucida measured by capillary suction: is zona hardening a mechanical phenomenon? J Exp Zool. 1988;245:206-219. 28. Schiewe MC, Araujo E, Asch RH, Balmaceda JP. Enzymatic characterization of zona pellucida hardening in human eggs and embryos. J Assist Reprod Gen. 1995;12:2-7.

98

References

29. Ducibella T, Kurasawa S, Rangarajan S, Kopf GS, Schultz RM. Precocious loss of cortical granules during mouse oocyte meiotic maturation and correlation with an egg- induced modification of the zona pellucida. Dev Biol. 1990;137:46-55. 30. Palmstierna M, Murkes D, Csemiczky G, Andersson O, Wramsby H. Zona pellucida thickness variation and occurrence of visible mononucleated blastomeres in preembryos are associated with a high pregnancy rate in IVF treatment. J Assist Reprod Genet. 1998;15:70-75. 31. Wassarman PM. Zona pellucida glycoproteins. J Biol Chem. 2008;283:24285-24289. 32. Ducibella T, Anderson E, Albertini DF, Aalberg J, Rangarajan S. Quantitative studies of changes in cortical granule number and distribution in the mouse oocyte during meiotic maturation. Dev Biol. 1988;130:184-197. 33. Ducibella T, Duffy P, Buetow J. Quantification and localization of cortical granules during oogenesis in the mouse. Biol Reprod. 1994;473:467-473. 34. Floehr J. Die Analyse von Fetuin-B bei idiopathisch infertilen Frauen. Master thesis. RWTH Aachen University. 2011. 35. Zegers-Hochschild F, Adamson GD, DeMouzon J, et al. The International Committee for Monitoring Assisted Reproductive Technology (ICMART) and the World Health Organization (WHO) revised glossary on ART terminology, 2009. Hum Reprod. 2009;24:2683-2687. 36. Thoma ME, Mclain AC, Louis JF, et al. The prevalence of infertility in the United States as estimated by the current duration approach and a traditional constructed approach. Fertil Steril. 2013;99:1324-1331. 37. Matzuk MM, Lamb DJ. Genetic dissection of mammalian fertility pathways. Nat Med. 2002;8:S41-S49. 38. Edwards RG, Steptoe PC. Birth after the reimplantation of a human embryo. Lancet. 1978;312:366. 39. European Society of Human Reproduction and Embryology. Annual meeting 2012. Focus Reprod. 2012;September:8. 40. DeMouzon J, Goossens V, Bhattacharya S, et al. Assisted reproductive technology in Europe, 2006: results generated from European registers by ESHRE. Hum Reprod. 2010;25:1851-1862. 41. Ferraretti AP, Goossens V, DeMouzon J, et al. Assisted reproductive technology in Europe, 2008: results generated from European registers by ESHRE. Hum Reprod. 2012;27:2571-2584.

99

References

42. Nyboe Andersen A, Goossens V, Ferraretti AP, et al. Assisted reproductive technology in Europe, 2004: results generated form European registers by ESHRE. Hum Reprod. 2008;23:756-771. 43. Foong SC, Fleetham JA, O’Keane JA, Scott SG, Tough SC, Greene CA. A prospective randomized trial of conventional in vitro fertilization versus intracytoplasmic sperm injection in unexplained infertility. J Assist Reprod Genet. 2006;23:137-140. 44. Bhattacharya S, Hamilton MPR, Shaaban M, et al. Conventional in-vitro fertilisation versus intracytoplasmic sperm injection for the treatment of non-male-factor infertility: A randomised controlled trial. Lancet. 2001;357:2075-2079. 45. Nyboe Andersen A, Carlsen E, Loft A. Trends in the use of intracytoplasmatic sperm injection marked variability between countries. Hum Reprod Update. 2008;14:593-604. 46. Johnson MH. Essential Reproduction. 6th ed. Malden, USA: Wiley-Blackwell; 2007. 47. Epifano O, Liang LF, Familari M, Moos MC, Dean J. Coordinate expression of the three zona pellucida genes during mouse oogenesis. Development. 1995;121:1947- 1956. 48. McNatty KP, Smith DM, Makris A, Osathanondh R, Ryan KJ. The microenvironment of the human antral follicle: interrelationships among the steroid levels in antral fluid, the population of granulosa cells, and the status of the oocyte in vivo and in vitro. J Clin Endocrinol Metab. 1979;49:851-860. 49. Schweigert FJ, Gericke B, Wolfram W, Kaisers U, Dudenhausen JW. Peptide and protein profiles in serum and follicular fluid of women undergoing IVF. Hum Reprod. 2006;21:2960-2968. 50. Andersen MM, Kroll J, Byskov AG, Faber M. Protein composition in the fluid of individual bovine follicles. J Reprod Fertil. 1976;48:109-118. 51. Spitzer D, Murach KF, Lottspeich F, Staudach A, Illmensee K. Different protein patterns derived from follicular fluid of mature and immature human follicles. Hum Reprod. 1996;11:798-807. 52. Angelucci S, Ciavardelli D, DiGiuseppe F, et al. Proteome analysis of human follicular fluid. Biochim Biophys Acta. 2006;1764:1775-1785. 53. Wang XN, Greenwald GS. Hypophysectomy of the cyclic mouse. II. Effects of follicle- stimulating hormone (FSH) and luteinizing hormone on folliculogenesis, FSH and human chorionic gonadotropin receptors, and steroidogenesis. Biol Reprod. 1993;48:595-605. 54. Dorrington JH, Moon YS, Armstrong DT. Estradiol-17beta biosynthesis in cultured

100

References

granulosa cells from hypophysectomized immature rats; stimulation by follicle- stimulating hormone. Endocrinology. 1975;97:1328-1331. 55. Hirshfield AN, Midgley AR. The role of FSH in the selection of large ovarian follicles in the rat. Biol Reprod. 1978;19:606-611. 56. Kumar TR, Wang Y, Matzuk MM. Follicle stimulating hormone is required for ovarian follicle maturation but not male fertility. Nat Genet. 1997;15:201-204. 57. Göritzlehner G, Lauritzen C, Göritzlehner U. Praktische Hormontherapie in Der Gynäkologie. 4th ed. Berlin, New York: Walter de Gruyter; 2003. 58. Pall M, Mikuni M, Mitsube K, Brannstrom M. Time-dependent ovulation inhibition of a selective progesterone-receptor antagonist (Org 31710) and effects on ovulatory mediators in the in vitro perfused rat ovary. Biol Reprod. 2000;63:1642-1647. 59. Lydon JP, DeMayo FJ, Conneely OM, O’Malley BW. Reproductive phenotpes of the progesterone receptor null mutant mouse. J Steroid Biochem Mol Biol. 1996;56:67-77. 60. Geist C, Harder U, Kriegerowski-Schröteler G, Stiefel A. Hebammenkunde, Lehrbuch Für Schwangerschaft, Geburt, Wochenbett Und Beruf. New York: Walter de Gruyter; 1995. 61. Müller WA, Hassel M. Entwicklungsbiologie Und Reproduktionsbiologie Des Menschen Und Bedeutender Modelorganismen. 5th ed. Berlin, Heidelberg: Springer Verlag; 2012. 62. Johnson S, Pion C, Jennings V. Current methods and attitudes of women towards contraception in Europe and America. Reprod Health. 2013;10:7. 63. Sturgis SH, Albright F. The mechanism of estrin therapy in the relief of dysmenorrhea. Endocrinology. 1940;26:68-72. 64. Bickenbach W, Paulicovics E. Hemmung der Follikelreifung durch Progesteron bei der Frau. Zbl Gynäk. 1944;68:153-157. 65. Lopez LM, Chen M, Mullins S, Curtis KM, Helmerhorst FM. Steroidal contraceptives and bone fractures in women: evidence from observational studies. Cochrane Database Syst Rev. 2015;7:CD009849. 66. Clemons M, Goss P. Estrogen and the risk of breast cancer. N Engl J Med. 2001;344:276-285. 67. Svendal G, Berk M, Pasco JA, Jacka FN, Lund A, Williams LJ. The use of hormonal contraceptive agents and mood disorders in women. J Affect Disord. 2012;140:92-96. 68. Tepper NK, Marchbanks PA, Curtis KM. Superficial venous disease and combined hormonal contraceptives: A systematic review. Contraception. 2016;94:275-279.

101

References

doi:10.1016/j.contraception.2015.03.010. 69. Birch Petersen K, Hvidman HW, Forman JL, et al. Ovarian reserve assessment in users of oral contraception seeking fertility advice on their reproductive lifespan. Hum Reprod. 2015;30:2364-2375. 70. Finer LB, Henshaw SK. Disparities in rates of unintended pregnancy in the United States, 1994 and 2001. Perspect Sex Reprod Health. 2006;38:90-96. 71. Yazdkhasti M, Pourreza A, Pirak A. Unintended pregnancy and its adverse social and economic consequences on health system: a narrative review article. Iran J Public Heal. 2015;44:12-21. 72. Shivers CA, Dudkiewicz AB, Franklin LE, Fussell EN. Inhibition of sperm-egg interaction by specific antibody. Science. 1972;178:1211-1213. 73. East J, Gulyas B, Dean J. Monoclonal antibodies to the murine zona pellucida protein with sperm receptor activity: effects on fertilization and early development. Dev Biol. 1985;273:268-273. 74. Lou Y, Ang J, Thai H, McElveen F, Tung KS. A zona pellucida 3 peptide vaccine induces antibodies and reversible infertility without ovarian pathology. J Immunol. 1995;155:2715-2720. 75. Greenhouse S, Castle PE, Dean J. Antibodies to human ZP3 induce reversible contraception in transgenic mice with “humanized” zonae pellucidae. Hum Reprod. 1999;14:593-600. 76. East IJ, Mattison DR, Dean J. Monoclonal antibodies to the major protein of the murine zona pellucida: effects on fertilization and early development. Dev Biol. 1984;104:49- 56. 77. Paterson M, Jennings A, Wilson MR, Aitken RJ. The contraceptive potential of ZP3 and ZP3 peptides in a primate model. J Reprod Immunol. 2002;53:99-107. 78. Aitken RJ, Baker MA, Doncel GF, Matzuk MM, Mauck CK, Harper MJK. As the world grows : contraception in the 21st century. J Clin Investig. 2008;118:1330-1343. 79. Bennett CF, Swayze EE. RNA targeting therapeutics: molecular mechanisms of antisense oligonucleotides as a therapeutic platform. Annu Rev Pharmacol Toxicol. 2010;50:259-293. 80. Monia BP, Lesnik EA, Gonzalez C, et al. Evaluation of 2’-modified oligonucleotides containing 2’-deoxy gaps as antisense inhibitors of gene expression. J Biol Chem. 1993;268:14514-14522. 81. Teplova M, Minasov G, Tereshko V, et al. Crystal structure and improved antisense

102

References

properties of 2’-O-(2-methoxyethyl)-RNA. Nat Struct Biol. 1999;6:535-539. 82. Stein CA, Subasinghe C, Shinozuka K, Cohen JS. Physicochemical properties of phosphorothioate oligodeoxynucleotides. Nucleic Acids Res. 1988;16:3209-3221. 83. Henry S, Stecker K, Brooks D, Monteith D, Conklin B, Bennett CF. Chemically modified oligonucleotides exhibit decreased immune stimulation in mice. J Pharmacol Exp Ther. 2000;292:468-479. 84. Geary RS, Yu RZ, Watanabe T, et al. Pharmacokinetics of a tumor necrosis factor- alpha phosphorothioate 2´-O-(2-methoxyethyl) modified antisense oligonucleotide: comparison across species. Drug Metab disposiion. 2003;31:1419-1428. 85. Geary RS. Antisense oligonucleotide pharmacokinetics and metabolism. Expert Opin Drug Metab Toxicol. 2009;5:381-391. 86. Graham MJ, Lee RG, Bell TA, et al. Antisense oligonucleotide inhibition of apolipoprotein c-III reduces plasma triglycerides in rodents, nonhuman primates, and humans. Circ Res. 2013;112:1479-1490. 87. Sewell KL, Geary RS, Baker BF, et al. Phase I trial of ISIS 104838, a 2´-methoxyethyl modified antisense oligonucleotide targeting tumor necrosis factor-α. J Pharmacol Exp Ther. 2002;303:1334-1343. 88. Kastelein JJP, Wedel MK, Baker BF, et al. Potent reduction of apolipoprotein B and low-density lipoprotein cholesterol by short-term administration of an antisense inhibitor of apolipoprotein B. Circulation. 2006;114:1729-1735. 89. Monteith DK, Geary RS, Leeds JM, Johnston J, Monia BP, Levin AA. Preclinical evaluation of the effects of a novel antisense compound targeting C-raf kinase in mice and monkeys. Toxicol Sci. 1998;46:365-375. 90. Yu RZ, Kim T, Hong A, Watanabe TA, Gaus HJ, Geary RS. Cross-species pharmacokinetic comparison from mouse to man of a second-generation antisense oligonucleotide, ISIS 301012, targeting human apolipoprotein B-100. 2007;35:460- 468. 91. Zanardi TA, Han S, Jeong EJ, et al. Pharmacodynamics and subchronic toxicity in mice and monkeys of ISIS 388626, a second-generation antisense oligonucleotide that targets human sodium glucose. Am Soc Pharmacol Exp Ther. 2012;343:489-496. 92. Geary RS, Wancewicz E, Matson J, et al. Effect of dose and plasma concentration on liver uptake and pharmacologic activity of a 2’-methoxyethyl modified chimeric antisense oligonucleotide targeting PTEN. Biochem Pharmacol. 2009;78:284-291. 93. Juliano R, Bauman J, Kang H, Ming X. Biological barriers to therapy with antisense

103

References

and siRNA oligonucletides. Mol Pharm. 2009;6:686-695. 94. Mouse genome sequencing consortium, Waterston RH, Lindblad-Toh K, et al. Initial sequencing and comparative analysis of the mouse genome. Nature. 2002;420:520-562. 95. Roy A, Matzuk MM. Deconstructing mammalian reproduction: Using knockouts to define fertility pathways. Reproduction. 2006;131:207-219. 96. Rankin TL, O’Brien M, Lee E, Wigglesworth K, Eppig J, Dean J. Defective zonae pellucidae in Zp2-null mice disrupt folliculogenesis, fertility and development. Development. 2001;128:1119-1126. 97. Rankin TL, Familari M, Lee E, et al. Mice homozygous for an insertional mutation in the Zp3 gene lack a zona pellucida and are infertile. Development. 1996;122:2903- 2910. 98. Rankin TL, Tong ZB, Castle PE, et al. Human ZP3 restores fertility in Zp3 null mice without affecting order-specific sperm binding. Development. 1998;125:2415-2424. 99. Baibakov B, Boggs NA, Yauger B, Baibakov G, Dean J. Human sperm bind to the N- terminal domain of ZP2 in humanized zonae pellucidae in transgenic mice. J Cell Biol. 2012;197:897-905. 100. Avella MA, Baibakov B, Dean J. A single domain of the ZP2 zona pellucida protein mediates gamete recognition in mice and humans. J Cell Biol. 2014;205:801-809. 101. Allen E. The oestrous cycle in the mouse. Am J Anat. 1922;30:297-371. 102. Bronson FH, Dagg CP, Snell GD. Reproduction. In: Green EL, Fahey EU, eds. Biology of the Laboratory Mouse. 2nd ed. New York: Dover Publications, Inc.; 2007. 103. The Jackson Laboratory. https://www.jax.org/. Accessed February 16, 2016. 104. Zurita E, Chagoyen M, Cantero M, et al. Genetic polymorphisms among C57BL/6 mouse inbred strains. Transgenic Res. 2011;20:481-489. 105. Freeman HC, Hugill A, Dear NT, Ashcroft FM, Cox RD. Deletion of nicotinamide nucleotide transhydrogenase: A new quantitive trait locus accounting for glucose intolerance in C57BL/6J mice. Diabetes. 2006;55:2153-2156. 106. Pettitt S, Liang Q, Rairdan X. Agouti C57BL/6N embryonic stem cells for mouse genetic resources. Nat Methods. 2009;6:493-495. 107. Bothe GWM, Bolivar VJ, Vedder MJ, Geistfeld JG. Genetic and behavioral differences among five inbred mouse strains commonly used in the production of transgenic and knockout mice. Genes, Brain Behav. 2004;3:149-157. 108. Bryant CD, Zhang NN, Sokoloff G, et al. Behavioral differences among C57BL/6 substrains: implications for transgenic and knockout studies. J Neurogenet.

104

References

2008;22:315-331. 109. Mulligan MK, Ponomarev I, Boehm SL, et al. Alcohol trait and transcriptional genomic analysis of C57BL/6 substrains. Genes, Brain Behav. 2008;7:677-689. 110. Philip VM, Sokoloff G, Ackert-Bicknell CL, et al. Genetic analysis in the collaborative cross breeding population. Genome Res. 2011;21:1223-1238. 111. Murray SA, Morgan JL, Kane C, et al. Mouse gestation length is genetically determined. PLoS One. 2010;5:1-7. 112. The Jackson Laboratory. Breeding performance survey, females of 35 commonly used strains of JAX Mice. MPD:Jax5. Mouse Phenome Database web site, The Jackson Laboratory, Bar Harbor, Maine USA. http://phenome.jax.org. Accessed April 18, 2016. 113. Runner MN, Ladman AJ. The time of ovulation and its diurnal regulation in the post- parturitional mouse. Anaomical Rec. 1950;108:343-361. 114. Fekete E. Observations on three functional tests in a high-tumor and a low-tumor strain of mice. Assoc Cancer Res. 1940;38:234-238. 115. Parkes AS. Estrous cycle of albino mouse. Proc R Soc London. 1926;100:151-170. 116. Voss R. Male accessory glands and the evolution of copulatory plugs in rodents. Occas Pap museum Zool Univ Michigan. 1979;689:1-27. 117. Byers SL, Payson SJ, Taft RA. Performance of ten inbred mouse strains following assisted reproductive technologies (ARTs). Theriogenology. 2006;65:1716-1726. 118. Weiskirchen R, Kneifel J, Weiskirchen S, van de Leur E, Kunz D, Gressner AM. Comparative evaluation of gene delivery devices in primary cultures of rat hepatic stellate cells and rat myofibroblasts. BMC Cell Biol. 2000;1. doi:10.1186/1471-2121-1- 4. 119. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature. 1970;227:680-685. 120. Baker BF, Lot SS, Condon TP, et al. 2’-O-(2-methoxy)ethyl-modified anti-intercellular adhesion molecule 1 (ICAM-1) oligonucleotides selectively increase the ICAM-1 mRNA level and inhibit formation of the ICAM-1 translation initiation complex in human umbilical vein endothelial cells. J Biol Chem. 1997;272:11994-12000. 121. Rasband WS. ImageJ. U.S. National Institutes of Health, Bethesda, Maryland, USA, http://imagej.nih.gov/ij/, 1997-2005. 122. Floehr J, Dietzel E, Neulen J, Rösing B, Weissenborn U, Jahnen-Dechent W. Association of high fetuin-B concentrations in serum with fertilization rate in IVF: a cross-sectional pilot study. Hum Reprod. 2016;31:630-637.

105

References

123. Dietzel E, Floehr J, van de Leur E, Weiskirchen R, Jahnen-Dechent W. Recombinant fetuin-B protein maintains high fertilization rate in cumulus cell-free mouse oocytes. Mol Hum Reprod. 2016:[Epub ahead of print]. 124. Sachdev M, Mandal A, Mulders S. Oocyte specific oolemmal SAS1B involved in sperm binding through intracrosomal SLLP1 during fertilization. Dev Biol. 2013;363:40-51. 125. The Jackson Laboratory. Body Weight. Available at: https://www.jax.org/jax-mice- and-services/strain-data-sheet-pages/body-weight-chart-000664. Accessed March 10, 2016. 126. Dell’Aquila ME, DeFelici M, Massari S, Maritato F, Minoia P. Effects of fetuin on zona pellucida hardening and fertilizability of equine oocytes matured in vitro. Biol Reprod. 1999;61:533-540. 127. Primakoff P, Myles DG. Penetration, adhesion, and fusion in mammalian sperm-egg interaction. Science. 2002;296:2183-2185. 128. DeFelici M, Siracusa G. “Spontaneous” hardening of the zona pellucida of mouse oocytes during in vitro culture. Gamete Res. 1982;6:107-113. 129. Zhang X, Rutledge J, Armstrong DT. Studies on zona hardening in rat oocytes that are matured in vitro in a serum-free medium. Mol Reprod Dev. 1991;28:292-296. 130. Legge M. Oocyte and zygote zona pellucida permeability to macromolecules. J Exp Zool. 1995;271:145-150. 131. Powers RW, Chen L, Russell PT, Larsen J. Gonadotropin-stimulated regulation barrier is mediated by nitric oxide of blood-follicle. Am Physiol Soc. 1995;269:E290-E298. 132. Hess KA, Chen L, Larsen WJ. The ovarian blood follicle barrier is both charge- and size-selective in mice. Biol Reprod. 1998;58:705-711. 133. Fullwood MJ, Liu MH, Pan YF, et al. An oestrogen receptor alpha-bound human chromatin interactome. Nature. 2010;462:58-64. 134. Murakami T, Walczak R, Caron S, et al. The farnesoid X receptor induces fetuin-B gene expression in human hepatocytes. Biochem J. 2007;407:461-469. 135. Zhu J, Jiang H, He RB, et al. Association between etiologic factors in infertile couples and fertilization failure in conventional in vitro fertilization cycles. Andrology. 2015;3:717-722. 136. Bódis J, Peti AM, Sulyok E, Kovács GL, Várnagy A. Serum and follicular fluid fetuin- A in women undergoing in vitro fertilization. Clin Chem Lab Med. 2014;52:1313-1318. 137. Schroeder AC, Schultz RM, Kopf GS, Taylor FR, Becker RB, Eppig JJ. Fetuin inhibits

106

References

zona pellucida hardening and conversion of ZP2 to ZP2f during spontaneous mouse oocyte maturation in vitro in the absence of serum. Biol Reprod. 1990;43:891-897. 138. George MA, Johnson MH. Use of fetal bovine serum substitutes for the protection of the mouse zona pellucida against hardening during cryoprotectant addition. Hum Reprod. 1993;8:1898-1900. 139. Kalab P, Schultz RM, Kopf GS. Modifications of the mouse zona pellucida during oocyte maturation: inhibitory effects of follicular fluid, fetuin, and alpha 2HS- glycoprotein. Biol Reprod. 1993;49:561-567. 140. Bugge TH, Kombrinck KW, Flick MJ, Daugherty CC, Danton MJS, Degen JL. Loss of fibrinogen rescues mice from the pleiotropic effects of plasminogen deficiency. Cell. 1996;87:709-719. 141. Bugge TH, Flick MJ, Daugherty CC, Degen JL. Plasminogen deficiency causes severe thrombosis but is compatible with development and reproduction. Gene Dev. 1995;9:794-807. 142. Bleil JD, Wassarman PM. Mammalian sperm-egg interaction: identification of a glycoprotein in mouse egg zonae pellucidae possessing receptor activity for sperm. Cell. 1980;20:873-882. 143. Avella MA, Xiong B, Dean J. The molecular basis of gamete recognition in mice and humans. Mol Hum Reprod. 2013;19:279-289. 144. Gahlay G, Gauthier L, Baibakov B, Epifano O, Dean J. Gamete recognition in mice depends on the cleavage status of an egg’s zona pellucida protein. Science. 2010;329:216-219. 145. Modliński JA. The role of the zona pellucida in the development of mouse eggs in vivo. J Embryol Exp Morphol. 1970;23:539-547. 146. Rankin TL, Talbot P, Lee E, Dean J. Abnormal zonae pellucidae in mice lacking ZP1 result in early embryonic loss. Development. 1999;126:3847-3855. 147. Winuthayanon W, Bernhardt ML, Padilla-Banks E, et al. Oviductal estrogen receptor α signaling prevents protease-mediated embryo death. Elife. 2015;4:e10453. doi:10.7554/e.Life.10453. 148. Sato K. Polyspermy - Preventing mechanisms in mouse eggs fertilized in vitro. J Exp Biol. 1975;9:425-431. 149. Stewart-Savage J, Bavister BD. A cell surface block to polyspermy occurs in golden hamster eggs. Dev Biol. 1988;128:150-157. 150. Meex RC, Hoy AJ, Morris A, et al. Fetuin B is a secreted hepatocyte factor linking

107

References

steatosis to impaired glucose metabolism. Cell Metab. 2015;22:1078-1089. 151. Henry SP, Denny KH, Templin MV, Yu RZ, Levin AA. Effects of an antisense oligonucleotide inhibitor of human ICAM-1 on fetal development in rabbits. Birth Defects Res B Dev Reprod Toxicol. 2004;71:368-373. 152. Henry SP, Denny KH, Templin MV, Yu RZ, Levin AA. Effects of human and murine antisense oligonucleotide inhibitors of ICAM-1 on reproductive performance, fetal development, and post-natal development in mice. Birth Defects Res Part B - Dev Reprod Toxicol. 2004;71:359-367. 153. Henry SP, Kim T-W, Kramer-Stickland K, Zanardi TA, Fey RA, Levin AA. Toxikologic properties of 2´-O-methoxyethyl chimeric antisense inhibitors. In: Crooke ST, ed. Antisense Drug Technology: Principles, Strategies, and Applications. 2nd ed. Boka Raton: CRC Press Taylor & Francis Group; 2008:352-353. 154. Kawamura K, Kawamura N, Kumagai J, Fukuda J, Tanaka T. Tumor necrosis factor regulation of apoptosis in mouse preimplantation embryos and its antagonism by transforming growth factor alpha/phosphatidylionsitol 3-kinase signaling system. Biol Reprod. 2007;76:611-618. 155. Clark DA, Lea RG, Podor T, Daya S, Banwatt D, Harley C. Cytokines determining the success or failure of pregnancy. Ann N Y Acad Sci. 1991;626:524-536. 156. Krishnan L, Guilbert LJ, Wegmann TG, Belosevic M, Mosman T. T helper 1 response against Leishmania major in pregnant C57BL/6 mice increases implantation failure and fetal resorptions. Correlation with increased IFN-gamma and TNF and reduced IL-10 production by placental cells. Am Assoc Immunol. 1996;156:653-662. 157. Driver SE, Robinson GS, Flanagan J, et al. Oligonucleotide-based inhibition of embryonic gene expression. Nat Biotechnol. 1999;17:1184-1187. 158. Gaudette MF, Hampikian G, Metelev V, Agrawal S, Crain WR. Effect on embryos of injection of phosphorothioate-modified oligonucleotides into pregnant mice. Antisense Res Dev. 1993;3:391-397. 159. Graham MJ, Crooke ST, Monteith DK, et al. In vivo distribution and metabolism of a phosphorothioate oligonucleotide within rat liver after intravenous administration. J Pharmacol Exp Ther. 1998;286:447-458. 160. Prakash TP, Graham MJ, Yu J, et al. Targeted delivery of antisense oligonucleotides to hepatocytes using triantennary N-acetyl galactosamine improves potency 10-fold in mice. Nucleic Acids Res. 2014;42:8796-8807. 161. Levin AA. A review of issues in the pharmacokinetics and toxicology of

108

References

phosphorothioate antisense oligonucleotides. Biochim Biophys Acta - Gene Struct Expr. 1999;1489:69-84. 162. Harenberg J, Marx S, Krejczy M, Wehling M. New anticoagulants - promising and failed developments. Br J Pharmacol. 2012;165:363-372. 163. Nicolaes G, Kulharia M, Voorberg J, et al. Rational design of small molecules targeting the C2 domain of coagulation factor VIII. Blood. 2014;123:113-120. 164. Pires E, Hlavin C, Macnamara E, et al. SAS1B protein shows temporal and spatial restriction to oocytes in several eutherian orders and initiates translation at the primary to secondary follicle transition. Dev Dyn. 2013;29:997-1003. 165. Quesada V, Sánchez LM, Alvarez J, López-Otín C. Identification and characterization of human and mouse ovastacin: a novel metalloproteinase similar to hatching enzymes from arthropods, birds, amphibians, and . J Biol Chem. 2004;279:26627-26634. 166. McAvey BA, Wortzman GB, Williams CJ, Evans JP. Involvement of calcium signaling and the actin cytoskeleton in the membrane block to polyspermy in mouse eggs. Biol Reprod. 2002;67:1342-1352. 167. Sengoku K, Tamate K, Horikawa M, Takaoka Y, Ishikawa M, Dukelow WR. Plasma membrane block to polyspermy in human oocytes and preimplantation embryos. J Reprod Fertil. 1995;105:85-90. 168. Jahnen-Dechent W, Heiss A, Schäfer C, Ketteler M. Fetuin-A regulation of calcified matrix metabolism. Circ Res. 2011;108:1494-1509.

109

110

Abbreviations

Abbreviations -mer derived from Greek, English = part 2`-MOE 2´-O-(2-methoxyethyl)ribonucleotides Ad5 adenovirus5 AHSG fetuin-A (genetic symbol, human) ALT alanine aminotransferase ART assisted reproductive techniques ASO antisense oligonucleotide AST aspartate aminotransferase ASTL ovastacin (genetic symbol, human) Astl ovastacin (genetic symbol, mouse) Astl-/- homozygote ovastacin deficient mouse Astl+/+ homozygote ovastacin wildtype mouse BSA bovine serum albumin Ca2+ calcium cDNA complementary deoxyribonucleic acid COC cumulus-oocyte complex CV coefficient of variation D1 domain 1 D2 domain 2 D3 domain 3 DKO double knockout DNA deoxyribonucleic acid dNTP deoxyribonucleotide triphosphates E2 estradiol ECLIA electrochemiluminescence immunoassay

EE2 17α-ethinyl estradiol ELISA enzyme-linked immunosorbent assay ERα estrogen receptor α FCS fetal calf serum FELASA Federation for Laboratory Animal Science Associations Fetub fetuin-B (genetic symbol, mouse) fetuB fetuin-B (protein) 111

Abbreviations

FETUB fetuin-B (genetic symbol, human) Fetub-/- homozygote fetuin-B deficient mouse Fetub+/- hemizygote fetuin-B mouse Fetub+/+ homozygote fetuin-B wildtype mouse FRET fluorescence resonance energy transfer FSH follicle stimulating hormone FXR farnesoid X receptor GAPDH glycerinaldehyd-3-phosphate dehydrogenase Gb gigabases GFP green fluorescence protein GN3 triantennary N-acetyl galactosamine GnRH gonadotropin releasing hormone HAT media hypoxanthine-aminopterin-thymidine medium hCG human chorionic gonadotropin His histidine HK heavy chain HRG histidine-rich glycoprotein (genetic symbol, human) HRP horseradish peroxidase HTF human tubal fluid

IC50 half maximal inhibitory concentration ICSI intracytoplasmatic sperm injection IgG immune globulin G IU international unit IVF in vitro fertilization kDa kilodaltons KNG kininogen (genetic symbol, human) LANUV Landesamt für Natur-, Umwelt-, und Verbraucherschutz LH luteinizing hormone LK light chain mA milliampere mAb monoclonal antibody MBCD methyl-β-cyclodextrin Mg2+ magnesium n.d. not determined

112

Abbreviations n.s. not significant Nnt nicotinamide nucleotide transhydrogenase P4 progesterone PEG polyethylene glycol rmfetuB recombinant mouse fetuin-B RNA ribonucleic acid RT-qPCR reverse transcription quantitative polymerase chain reaction SDS-PAGE sodium dodecyl sulfate polyacrylamide gel Ser serum shRNA short hairpin RNA SLLP1 sperm lysozyme-like protein 1 SNP single nucleotide polymorphism Taq Termus aquaticus TBS TRIS buffered saline TMB tetramethylbenzidine TNFα tumor necrosis factor α TRIS tris(hydroxymethyl)aminomethane TYH Toyoda Yokoyama Hoshi US United States WT wildtype ZP zona pellucida ZP1 zona pellucida protein 1 Zp2 mouse zona pellucida 2 gene ZP2 zona pellucida protein 2 Zp3 mouse zona pellucida 3 gene ZP3 zona pellucida protein 3 ZP4 zona pellucida protein 4

113

114

Figures

Figures Figure 1. Structural comparison of type 3 cystatins 11 Figure 2. Physiological role of fetuin-B in fertilization 13 Figure 3. Scheme of the murine zona pellucida 13 Figure 4. Mouse and human fetuin-B alignment 15 Figure 5. Schematic overview of folliculogenesis 17 Figure 6. Hormonal menstrual cycle regulation, schematic diagram 19 Figure 7. 20-mer fetuin-B antisense oligonucleotide sequence, a second-generation antisense oligonucleotide 22 Figure 8. Antisense oligonucleotide mechanism 23 Figure 9. Serum fetuin-B level variation over the course of one month 45 Figure 10. Serum fetuin-B in women on contraceptive medication 46 Figure 11. Matched serum fetuin-B and serum progesterone, estradiol and LH in untreated women 47 Figure 12. Association of serum and follicular fluid fetuin-B 48 Figure 13. Association of serum fetuin-B and serum estradiol and LH in women undergoing hormonal ovarian stimulation 49 Figure 14. Association of serum fetuin-B and endogenous estradiol during pregnancy 50 Figure 15. Association of serum fetuin-B and fertilization rate in IVF 51 Figure 16. Human fetuin-B inhibited ovastacin 52 Figure 17. Purification of recombinant mouse fetuin-B 53 Figure 18. Evaluation of antigen binding potential in serum of immunized Fetub-/- mice 55 Figure 19. Evaluation of antigen binding potential in supernatant from monoclonal antibody-producing hybridoma cells after the second cloning 58 Figure 20. Evaluation of the supernatant of hybridoma cells after the second cloning by native and reducing Western blot 59 Figure 21. Antigen detection by the supernatant from monoclonal antibody-producing hybridoma cells, before and after antibody purification 60 Figure 22. Evaluation of monoclonal antibodies 20-4 and 27-10 by native Western blot 61 Figure 23. Evaluation of monoclonal antibodies 20-4 and 27-10 by reducing Western blot 61 Figure 24. Immunoprecipitation of mouse fetuin-B by monoclonal antibodies 63 Figure 25. Fetuin-B/ovastacin double deficiency restores fertility of fetuin-B deficient female mice 64 Figure 26. Offspring of double deficient females had normal weight 65 Figure 27. In vitro fertilization rates of double deficient (DKO) and wildtype (WT) females 66 Figure 28. Serum fetuin-B down-regulation by fetuin-B ASO 67 115

Figures

Figure 29. Serum fetuin-B of PBS and ASO-treated females 69 Figure 30. Quantification of serum fetuin-B by recombinant mouse fetuin-B as standard 70 Figure 31. Fetuin-B ASO-mediated down-regulation of serum fetuin-B causes infertility 71 Figure 32. Microscopic analysis of an unstable vaginal plug during fetuin-B ASO treatment 73 Figure 33. Fetuin-B ASO led to reduced fetuin-B mRNA synthesis and increased aminotransferase activity 75 Figure 34. Weight development of PBS and fetuin-B ASO-treated females 76 Figure 35. Fetuin-B down-regulation is reversible and did not affected the offspring 78 Figure 36. Serum fetuin-B is associated with fertility 80 Figure 37. Alanine (ALT) and aspartate (AST) aminotransferase activity following PBS, control oligonucleotide or fetuin-B ASO treatment 81 Figure 38. Interaction of fetuin-B and ovastacin during fertilization 87

116

Tables

Tables Table 1. Reproductive data of C57BL/6J mice 25 Table 2. Fetub and Astl genotyping PCR components 37 Table 3. Fetub and Astl genotyping PCR conditions 37 Table 4. Human tubular fluid (HTF) medium composition 39 Table 5. Modified Krebs-Ringer bicarbonate medium composition 40 Table 6. Screening of primary culture cells following the selection phase 56 Table 7. Cross-reactivity testing following the first cloning step 56 Table 8. Cross-reactivity testing following the second cloning step 57 Table 9. Evaluation of antigen detection by the clones 15-1, 20-4 and 27-10 57 Table 10. Evaluation of antigen detection by purified monoclonal antibodies 20-4 and 27-10 62 Table 11. Mating behavior of fetuin-B ASO-treated females (repetitive 50 mg/kg boli) 68 Table 12. Mating behavior of PBS and ASO-treated mice (repetitive 100 mg/kg boli) 72 Table 13. Mating behavior of fetuin-B hemizygous (+/-) and fetuin-B deficient (-/-) females 73 Table 14. Reproductive behavior following ASO treatment cessation 77

117

118

Acknowledgement

Acknowledgment

Zuerst möchte ich mich bei meinem Betreuer Prof. Dr. Willi Jahnen-Dechent für sein Vertrauen in das Projekt und meine Arbeit bedanken - für die Möglichkeit mit dem Projekt zu wachsen. Danke für die zahlreichen Diskussionen, die Ideen, die daraus entstanden sind und die Unterstützung während der gesamten Zeit.

Danke auch an Prof. Dr. Marc Spehr für sein Interesse an dem Projekt und seine Unterstützung zur Generierung der Fetuin-B/Ovastacin doppelt defizienten Mäuse, sowie für seine Bereitschaft als Zweitprüfer.

Zudem gilt mein Dank: dem Team der Frauenklinik für Gynäkologische Endokrinologie und Reproduktionsmedizin, dort vor allem Dr. Ute Weißenborn, Dr. med. Benjamin Rösing und Prof. Dr. med. Joseph Neulen, für die gute Zusammenarbeit, für den regen Austausch und die Unterstützung bei allen klinischen Fragestellungen. many thanks to Alfred Chappell for his great support and helpful hints during the ASO study.

Christian Spoden und Annika Peschel aus dem Institut für Versuchstierkunde. Die unkomplizierten Absprachen haben die Arbeit sehr erleichtert.

Prof. Dr. Ralf Weiskirchen und Eddy van de Leur für die Herstellung der rekombinanten Proteine.

Hagen Westphahl und Konstantin Karmilin für ihre Arbeit rund um Ovastacin und die gute Kooperation zwischen den Arbeitsgruppen.

Dagmar Dechent für ihre Hilfe bei statistischen Fragestellungen. insbesondere Eileen Dietzel. Es ist unbezahlbar, dass es jemanden gibt, mit dem man sich austauschen kann und der einen unterstützt. Es hat einfach Spaß gemacht gemeinsam so viel Neues zu entdecken. bei Marietta Herrmann für ihr Verständnis und ihre konstruktive Kritik. Sie war für mich immer hilfreich und motivierend. Beide haben meine Arbeitsweise wohl maßgeblich mitgeprägt. Danke euch beiden dafür!

Carlo Schmitz für die gute Zusammenarbeit und seine Unterstützung bei der Generierung der Fetuin-B/Ovastacin doppelt defizienten Mäuse. Hier ebenso Sina Köppert. der gesamten Arbeitsgruppe für die Wohlfühlatmosphäre, vor allem Steffen Gräber und Anne Babler. 119

Acknowledgement

Niklas, du wusste vielleicht nicht, was du tust, aber dein Lachen ist großartig und macht jedes Ärgernis nichtig.

Danke Andreas, für deinen Rückhalt und deine Geduld. Danke, dass du mich zum Lachen bringst!

Danke Melle. Für jeden Schritt und dein wortloses Verständnis.

Besonderer Dank gilt meinen liebevollen Eltern. Danke für eure Unterstützung; für all das, was ihr mir mitgegeben habt! Und meinen Geschwistern Katharina, Tilman und Kilian. Es ist ein Geschenk eure kleine Schwester zu sein!

120