Atm deficient zebrafish model reveals conservation of the tumour suppressor function and a role in fertility

Abstract

Biallelic loss-of-function variants in ATM (Ataxia Telangiectasia Mutated) cause Ataxia Telangiectasia (AT), a rare disorder associated with cerebellar degeneration and ataxia, cancer predisposition, infertility, growth retardation, etc. ATM is a phosphoinositide 3-kinase-related kinase (PIKK) with a role in DNA repair and maintenance of genome stability. Studying a multisystem genetic disease like AT requires animal models to ascertain its pathogenesis at the level of tissues, organs and the organism. Due to its small size, cheap maintenance, large progeny, rapid development and initial transparency, zebrafish (Danio rerio) is an increasingly popular vertebrate model organism, suitable for genetic modifications and large-scale in vivo therapeutic screens as embryos are chemically permeable to small compounds. Currently, no zebrafish model for AT exists. 1 We generated atm knock-outs through CRIPSR-Cas9 mutagenesis. We show that atm conserved its function as a tumour suppressor gene and is involved in gametogenesis and fertility. Therefore, this mutant is of great value for further studies investigating the role of atm in reproduction and tumorigenesis. Zebrafish atm evolved to a single copy after the teleost genome duplication, which appears distinctive for DNA damage repair genes. 2 Human ATM and zebrafish atm both comprise 62 coding exons with high amino acid conservation in functional domains (Fig. 1A and Table S1). Therefore, ATM functions may be conserved in zebrafish and atm deficiency might recapitulate human disease. Figure 1 Characteristics of Atm deficient zebrafish. (A) Zebrafish and human Atm/ATM structure. atm mutations introduced by CRISPR-Cas9 mutagenesis are indicated. Protein domains of human ATM (3056 aa) and zebrafish Atm (3091 aa) are highly conserved. The encoded proteins display an overall AA sequence identity of 54% with a higher conservation for the important functional protein domains (yellow = TAN (Tel1/ATM N-terminal motif: LxxxKxxE/DRxxxL); black stripe = c-ABL; orange = FAT (FRAP-ATM-TRRAP) domain; blue = PI3K (phosphoinositide 3-kinase-like) domain; green = FATC (FAT C-terminal). These domains display a homology of 77, 78, 60, 82 and 94% respectively, with also Ser1981, an important autophosphorylated site, being preserved at AA position 2012 in zebrafish (B) atm deficient zebrafish developed exclusively as males (P < 10−5, Fisher's exact test). Mendelian inheritance laws rule out embryonic lethality. Black line = 1 cm. (C) Histology of zebrafish gonads. Zebrafish atm wild types and heterozygotes have fully functional gonads. Oocytes in all stages of maturation are present, while testes have adult sperm in their lumen. Zebrafish atm mutants do not develop ovaria. Zebrafish atm testis are devoid of spermatids and spermatozoa. Red L = lumen, I = Stage I oocytes, II = Stage II oocytes, III = Stage III oocytes. (D) Wild-type testis has spermatogonia, spermatocytes, spermatids and spermatozoa. In contrast, atm-mutants lack spermatids and spermatozoa in their lumen. Spermatocytes are arrested in the bouquet phase of meiosis, and some spermatocytes appear pyknotic. sg: spermatogonia, sc: spermatocytes, st: spermatids, sz: spermatozoa, pk: pyknosis. (E) Caspase-3 immunostaining of testis clearly shows cleaved caspase-3 positive apoptotic cells in atm-mutants. Black arrows indicate cells undergoing apoptosis. atm cmg33/cmg34 ;tp53 zdf1/zdf1 mutant zebrafish are prone to develop tumours. (F) Age of tumour onset is significantly lower in atm cmg33/cmg34 ;tp53 zdf1/zdf1 zebrafish compared to atm +/- ;tp53 zdf1/zdf1 and atm +/+ ;tp53 zdf1/zdf1 zebrafish. ∗: P < 10−3, ∗∗: P < 10−4. Unpaired t-test was used. Error bars indicate 95% Confidence Interval (CI). In GI-III fish with protruding tumours are shown (inset: tumour location indicated with red arrow; main figure: showing histopathological image of tumour (G.I) Malignant peripheral nerve sheath tumour (MPNST) in the eye of a double mutant zebrafish. The tail of this zebrafish was deformed since early development (G.II) undifferentiated spindle cell sarcoma (US) growing in the tail (G.III) MPNST in the coelom invading the intestinal lining. Fig. 1 Through CRISPR/Cas9 mutagenesis we generated two heterozygous atm knock-out mutant zebrafish lines, with premature stop codons in exon 4 (atm cmg34) and 9 (atm cmg33) (Fig. 1A; Fig. S1 and Table S2). For all experiments, these lines were in-crossed to obtain compound heterozygous mutants (atm cmg33/cmg34, “atm-mutants”), avoiding potential homozygosity of CRISPR-induced off-target mutations. Atm deficiency is not embryonically lethal: offspring of atm cmg33/+ xatm cmg34/+ (n = 77) displayed 27% wild types, 51% heterozygotes and 22% mutants (Fig. 1B), in agreement with Mendelian inheritance (Chi 2  = 0,80). Equal rates of males:females were found for wild types and heterozygotes; however, atm cmg33/cmg34 fish developed exclusively as males. This all-male phenotype was found over multiple crosses and generations of atm-mutants. Atm-mutant males were able to induce egg laying in wild-type females, but eggs remained unfertilized (Fig. S2). H&E (hematoxylin and eosin) stained histological sections of gonads displayed male testes in all atm-mutants (Fig. 1C) but lacked spermatids and adult spermatozoa (Fig. 1D), while wild-type testes contained cells at all stages of differentiation during spermatogenesis. Additionally, testes of mutants displayed large groups of spermatocytes in the bouquet phase of meiosis and some spermatocytes appeared pyknotic, suggesting apoptosis. Apoptosis was confirmed by Caspase-3 immunostaining and TUNEL labelling (Fig. 1E; Fig. S3) and indicates that zebrafish atm is essential in spermatogenesis. Zebrafish lack identifiable heteromorphic sex chromosomes. The current model hypothesizes that zebrafish sex is partially determined by the number of primordial germ cells (PGC), with reduced numbers causing zebrafish to become males. The Fanconi anaemia (FA) pathway is required for PGC to survive and thus allow female sex to occur. 2 ATM is attributed multiple roles in homologous recombination (HR), essential for repair of crosslinks through the FA pathway. However, ATM is also implicated in control of meiotic double strand break (DSB) formation. Absence of functional ATM could lead to over-activation of SPO11-mediated DSBs, ultimately leading to apoptosis. 3 The all-male phenotype suggests that PGC of juvenile zebrafish require atm for survival. Next, we investigated if apoptosis of PGC causes the all-male phenotype in atm-mutants by disabling p53-mediated apoptosis. Hereto we generated a double knock-out atm cmg33/cmg34 ;tp53 zdf1/zdf1 model. tp53 zdf1/zdf1 mutants displayed equal rates of males:females, while atm-mutants with a functional tp53 allele developed exclusively as males (Table S3). However, atm cmg33/cmg34 ;tp53 zdf1/zdf1 males could not fertilize eggs of wild-type females (Fig. S4A). Histology confirmed that spermatids and spermatozoa were absent in these zebrafish, similar to atm cmg33/cmg34 ; tp53 +/+ zebrafish (Fig. S4B). atm cmg33/cmg34 ;tp53 zdf1/zdf1 females produced some fertilized eggs, but oocytes likely acquired genomic aberrations, as all embryos died within 1 day post fertilization (Fig. S4C). Histology of the ovaries appeared normal (Fig. S4D). As atm +/+ ;tp53 zdf1/zdf1 and atm +/- ;tp53 zdf1/zdf1 females were able to produce healthy offspring, we conclude that atm is essential for the viability of zebrafish oocytes. Therefore, atm plays a role in zebrafish sex development and fertility. Infertility has been reported in AT individuals and multiple Atm-deficient animal models. Another key characteristic of AT patients is their increased cancer risk (30%–40% lifetime risk). We sectioned 12 atm-mutants of 16 months old but could not find tumours/neoplasia. This is not surprising, since knocking out tumour suppressors in zebrafish does not cause the same increased cancer risk as in humans: TP53 and BRCA2 germline mutations confer a 70%–100% cancer risk in humans. In zebrafish, tp53 deficiency caused a 40% cancer risk, 4 and only 30% of brca2-deficient zebrafish displayed testicular neoplasia at 16 months. 5 To accelerate tumour formation, we investigated atm-deficiency in a tp53-mutated background. atm cmg33/cmg34 ;tp53 zdf1/zdf1 zebrafish developed tumours much faster (average: 9.0 months) compared to atm +/- ;tp53 zdf1/zdf1 (average: 14,1 months, p < 10−4) or atm +/+ ;tp53 zdf1/zdf1 zebrafish (average: 13,4 months, p < 1 0−3) (Fig. 1F and Table S4), suggesting that atm conserved its tumour suppressor role and that atm deficiency accelerated tumorigenesis in tp53-mutant zebrafish. Malignant peripheral nerve sheath tumours (MPNST), undifferentiated spindle cell sarcomas (US) and seminomas were found (Fig. 1G), which are standard for tp53 zdf1 . AT patients regularly develop leukaemia, which we did not observe in atm-mutants. Similarly, brca2 −/− zebrafish do not present with haematological tumours while Fanconi anaemia patients develop such malignancies. Studying leukaemia in atm-mutants may require transgenic lines containing oncogenic drivers or another type of tp53 variant: tp53 zdf1 contains a missense variant and MPNSTs are typically found, while tp53 del/del harbours a deletion and showed a high proportion of leukaemia. 4 The ataxia-phenotype of AT patients was not reproduced in zebrafish atm-mutants, based on swim movements in a rotating chamber at the age of 6 or 12 months (Fig. S5A). This is in contrast to some murine, rat and porcine models. We cannot rule out that degenerative phenotypes only manifest at later ages. Alternatively, atm-deficient zebrafish may display a subtle phenotype, requiring more sophisticated methods, like studying the optokinetic response, as AT patients have trouble reading due to impaired coordination of the eye movement. The cerebellum of AT patients displays neurodegeneration, hallmarked by loss of Purkinje cells. Histological analysis of the cerebellum in aged atm-mutants revealed a normal Purkinje cell count and molecular layer thickness (Fig. S6A). Also in murine models, histology was unable to show abnormal cerebellar architecture. A higher complexity of the brain may be needed for atm to play a major role in neurodegeneration. To come to definite conclusions additional tests with larger sample sizes and zebrafish of different ages may be required. In contrast to human AT patients, growth retardation was not displayed in adult (12 months old) atm-deficient zebrafish compared to age-matched wild-type control males (Fig. S7). Generally, zebrafish with loss of DNA damage response (DDR) genes lack this phenotype. Zebrafish may suffer less from dysphagia, which in children leads to reduced caloric intake and compromises growth. Other explanations could be a role of modifier genes, or a lack of environmental challenges in laboratory settings. We cannot fully exclude residual functionality of the mutant protein as suitable Atm antibodies to check for residual protein were not available. In conclusion, despite the absence of a clear neurodegenerative phenotype, we show that zebrafish atm is required for sex-determination and fertility and acts as a tumour suppressor gene. The model could be useful to discern the role of this kinase in genome stability maintenance and tumour suppression and may be used to discover novel compounds targeting ATM deficiency. Conflict of interests Authors declare no conflict of interests. Funding This work was supported by ‘Kom op Tegen Kanker – Emmanuel Van der Schueren’ (No. 365M02318) and by ‘UGent- Bijzonder OnderzoeksFonds’ (No. BOF15 GOA/011).