Analysis of Potential Non-Canonical or Alternate STAT5 Functions in Immune Development and Growth

Background : Signal transducer and activator of transcription (STAT) proteins play key roles in development, growth, and homeostasis. These roles have principally been assigned to their “canonical” function as inducible transcriptional activators acting downstream of cytokines and other factors. However, variant “non-canonical” functions have also been identified. The potential in vivo role for non-canonical STAT functions was investigated in the zebrafish model. Methods : Two zebrafish Stat5.1 mutants were generated using CRISPR/Cas9 that should impact canonical functionality: one with a deleted transactivation domain ( ∆ TAD) and another with a disrupted tyrosine motif ( ∆ TM). Immune cell development, growth, and adiposity of these Stat5.1 mutants were assessed in comparison to a Stat5.1 knockout (KO) mutant in which both canonical and non-canonical functions were ablated. Results : Both the ∆ TAD and ∆ TM mutants showed significantly reduced embryonic T lymphopoiesis, similar to the KO mutant. Additionally, adult ∆ TAD and ∆ TM mutants displayed a decrease in T cell markers in the kidney, but not as severe as the KO, which also showed T cell disruption in the spleen. Severe growth deficiency and increased adiposity were observed in all mutants, but ∆ TAD showed a more modest growth defect whereas ∆ TM exhibited more profound impacts on both growth and adiposity, suggesting additional gain-of-function activity. Conclusions : These results indicate that canonical Stat5.1 plays a major role in T cell development and growth throughout the lifespan and non-canonical Stat5.1 functions also contribute to aspects of adult T lymphocyte development and growth, with alternate functions impacting growth and adiposity.


Introduction
Mammals possess seven signal transducer and activator of transcription (STAT) proteins: STAT1, STAT2, STAT3, STAT4, STAT5A, STAT5B, and STAT6 [1]. The myriad roles of STATs across development, growth, and homeostasis have been largely attributed to their so-called "canonical" function as inducible transcription factors downstream of cytokines and other factors [2,3]. In this modality, ligation of cytokine or other receptors triggers the tyrosine phosphorylation of inactive unphosphorylated STAT (uSTAT) monomers in the cytoplasm via associated kinases such as JAKs, allowing dimerization of phosphorylated STAT (pSTAT) molecules through reciprocal interactions between a phosphotyrosine-containing motif on one STAT protein with the SH2 domain on another. The subsequent nuclear translocation of these activated STAT dimers enables them to activate the transcription of target genes, impacting key cell processes, notably including differentiation, survival, proliferation, and activation [4][5][6]. However, STATs have also been demonstrated to regulate critical cell processes through alternative "non-canonical" functions [7].
Mammalian STAT5A and STAT5B proteins share more than 90% amino acid similarity and have both unique as well as overlapping functions [8]. Both STAT5A and STAT5B contribute to immune regulation, although STAT5B appears to play the dominant role, presumably via canonical signaling downstream of various interleukin (IL) receptors [9]. STAT5A is a master regulator of mammary gland development and lactogenesis attributed to a canonical signaling function downstream of prolactin [10], whereas STAT5B facilitates the regulation of growth and metabolism assigned to canonical growth hormone (GH) signaling [11][12][13]. Non-canonical functions of STAT5 proteins have also been reported. These include gene repression by pSTAT5B during oncogenesis [14], uSTAT5B participating in the regulation of megakaryocyte differentiation [15], and non-nuclear roles of STAT5A/B in maintaining the structural integrity of sub-cellular organelles such as the endoplasmic reticulum and Golgi body [16]. Various STAT5B mutations have been associated with disrupted growth and immunity but with differential clinical severity depending on the site of mutation, further suggesting the potential for non-canonical functions [12,[17][18][19].
Zebrafish possess two STAT5 proteins, Stat5.1 and Stat5.2 [20]. Stat5.1 displays the closest similarity with mammalian STAT5B and has been shown to contribute to lymphopoiesis throughout the life course [21], as well as the regulation of growth and sex-specific adiposity [21,22]. However, these publications used loss-of-function (LOF) knockout (KO) mutant models that would abrogate both canonical and non-canonical functions. In this study, two Stat5.1 mutants were generated using CRISPR/Cas9-based mutagenesis to examine potential non-canonical or alternate functions. The impact of these Stat5.1 mutants on lymphopoiesis, growth, and adiposity was examined in direct comparison to a Stat5.1 KO mutant. The results provided evidence of predominantly canonical functions but also suggested that non-canonical and alternate Stat5.1 functions also exist.

Zebrafish Husbandry
Zebrafish were maintained using standard husbandry practices [23] following national guidelines for their care and use. All studies were approved by the Deakin University Animal Ethics Committee.

Generation of Stat5.1 Mutants
A single guide RNA (sgRNA) targeting the zebrafish stat5.1 gene was designed using ZiFiTargetor v4.2 [24] and CHOPCHOP v3 [25], which had a minimum of three mismatches with potential off-target sites, all on other chromosomes (Supplementary Table 1). The sgRNA template was cloned into pDR274 and sgRNA synthesized by in vitro transcription using a MEGAscript T7 Transcription Kit (#AM1354, Thermofisher Scientific, Australia), purified using a MegaClear Kit (#AM1908, Thermofisher Scientific, Australia), and quantified. One cell stage wildtype (WT) zebrafish embryos were injected with~1 nL of 100 ng/µL sgRNA, 200 ng/µL TrueCut Cas9 protein v2 (#A36498, Thermofisher Scientific, Australia), and 0.4% (w/v) phenol red. Founder fish were raised to adulthood, out-crossed with WT fish, and screened for the presence of mutations. Confirmed heterozygous fish were further outcrossed to minimize potential off-target mutations, with heterozygous progeny in-crossed to obtain homozygous mutant lines.

Genomic DNA Analysis
Genomic DNA from adult fin clips and whole embryos was isolated with QuickExtract DNA Extraction Solution (#QE09050, Gene Target Solution, Australia) following the manufacturer's instructions. This was subjected to a polymerase chain reaction (PCR) for high-resolution melt (HRM) analysis with the stat5.1-specific primers 5 ′ -GCGAGGGTGTGAAAAACAGT and 5 ′ -CCCTTTCTGT CTTCCTGTCCA or gel electrophoresis with the alternate primers 5 ′ -GTGGTAAAAATGTGTGGATGAACTCTG and 5 ′ -GTCTCATGTATCCAAGGCAACTCG to identify potential mutants, which were confirmed by Sanger sequencing at the Australian Genome Research Facility.

Whole-Mount in Situ Hybridization
Embryos were collected and raised in egg water with 0.0003% (w/v) phenylthiourea (PTU) for 5 days and then fixed in 4% paraformaldehyde/phosphate-buffered saline (PFA/PBS) overnight at 4°C, followed by 100% methanol at -20°C. Whole-mount in situ hybridization (WISH) was performed on rehydrated embryos according to a published protocol [26]. Embryos and juvenile fish were imaged with a MVX10 monozoom microscope using a 1 × MVXPlan Apochromat lens (NA = 0.25) and DP74 camera (#N5667400, Olympus, Tokyo, Japan), with the images captured on CellSens Dimension 1.6 software (Olympus, Tokyo, Japan), and areas of WISH staining determined using ImageJ software (v1.x, National Institutes of Health (NIH), Bethesda, MD, USA).

Body Measurements
Embryos and juvenile fish were imaged as described above and the length was determined using ImageJ software. Adult fish were imaged using an iPhone 13 pro max, wide camera-26 mm f1.5 12 MP, 3024 × 4032, 1.5 µm pixel size (Apple Inc., Cupertino, CA, USA) next to a ruler, with standard length measured from snout to caudal peduncle [27]. Following removal of eggs adult female zebrafish were weighed and subjected to total lipid extraction using 8:4:3 (v/v) chloroform:methanol:NaCl (0.03%), incubating the organic phase at 65°C in a fumehood until dry to determine total lipid as described [28], which was expressed as a percentage of body weight.

Statistical Analysis
Statistical analysis was performed with GraphPad Prism v8.0.0 for Windows (GraphPad Software, San Diego, CA, USA) using One-way ANOVA with Tukey's multiple comparison test.

Generation of Zebrafish Stat5.1 Mutants
Zebrafish Stat5.1 is highly homologous to human STAT5B, both at the level of protein domains as well as individual amino acid residues [20], including the region spanning the SH2 domain, the tyrosine motif (TM) phosphorylated in response to the external stimuli [32], and the transactivation domain (TAD) (Fig. 1A). The region adjacent to the TM was targeted using CRISPR-Cas9 genome editing [33,34] with a specific sgRNA (Fig. 1B). Two different mutant alleles were generated, which were out-crossed twice and bred to homozygosity and sequenced (Fig. 1C): mdu033, a 4 bp deletion causing a frameshift that results in the complete deletion of the TAD (∆TAD) and mdu032, a 12 bp in-frame deletion resulting in the removal of 4 amino acids in the TM but with other sequences intact. These mutants were subsequently compared to wild-type zebrafish and a previously generated knockout Stat5.1 KO mutant [21] (Fig. 1D).

Impact of Stat5.1 Mutations on Lymphopoiesis
Mammalian STAT5 proteins play important roles in the development of lymphoid cells, particularly STAT5B [35]. We have recently shown that zebrafish Stat5.1 KO mutants exhibited severe disruption in T lymphopoiesis throughout the life course, as well as, other lymphoid defects [21]. Thus, we compared the impact of the other Stat5.1 mutants on this aspect of biology.
Embryonic lymphopoiesis was investigated using WISH on 5 dpf embryos using specific T cell markers: rag1 for early T cells [36] and lck for mature T cells [37,38]. KO embryos showed a significant reduction in the area of rag1 staining in comparison to WT, as described [21]. Both ∆TAD and ∆TM mutants showed a comparable significant reduction in the area of rag1 staining, similar to the KO ( Fig. 2A,C). Likewise, the area of lck staining was also reduced in both ∆TAD and ∆TM mutants compared to the WT to a similar extent as seen with the KO (Fig. 2B,D).
Adult lymphopoiesis was examined in the kidney marrow, a primary lymphoid organ that plays an equivalent role to mammalian bone marrow [39,40], and the spleen, a secondary lymphoid organ [41]. Expression of four different lymphocyte-specific markers was analyzed: tcra (T cells) [42], cd8 (cytotoxic T cells) [43], ighm (B cells) [44], and nkld (NK cells) [45]. In the kidney, expression of tcra and cd8 were both significantly reduced in the KO compared to the WT, as described [21]. Expression of cd8 was also significantly reduced in the ∆TAD and ∆TM mutants (Fig. 2E,F), although the reduction was not statistically significant for tcra in ∆TM and was more subtle than observed in the KO. No significant changes in the expression of ighm or nkld were observed in any of the mutants (Fig. 2E,F). In the spleen, the KO, but not the other mutants, exhibited significantly reduced cd8 expression compared to the WT, with the KO also significantly lower than the ∆TAD. The expression of ighm was upregulated in the KO compared to the WT, but remained unaltered in ∆TAD and displayed a significant reduction in ∆TM compared to the KO. The relative expression of tcra and nkld remained unchanged in all mutants compared to the WT (Fig. 2F).

Impact of Stat5.1 Mutations on Growth and Adiposity
Mammalian STAT5B/Stat5b plays a crucial role in the regulation of growth and adiposity [46]. Moreover, zebrafish Stat5.1 KO mutants have also been found to exhibit reduced growth but increased adiposity along with dysregulation of growth and lipid metabolism genes [21,22]. During all developmental stages and into adulthood for both sexes, the ∆TAD and ∆TM mutants displayed a significant reduction in length compared to the WT (Fig. 3A-D,  F-H), as well as weight in adults (Fig. 3I,J). However, there were subtle differences. Both the ∆TAD and ∆TM mutants were slightly larger than the KO at the embryonic stage (Fig. 3A,C). The trend continued for the ∆TAD mutant, and although not statistically significant at the juvenile stage    (Fig. 3B,D), it was for adult males but not females (Fig. 3F-J). In contrast, the ∆TM mutant was significantly smaller compared to the KO at the juvenile stage (Fig. 3B,D), as well as in adulthood of both sexes (Fig. 3F-J). Total lipid content was significantly increased in all Stat5.1 mutants compared to the WT, with the ∆TM mutant significantly higher than the KO (Fig. 3K).
To investigate the molecular underpinnings of the growth deficiency and enhanced adiposity, 28 dpf juvenile fish were analyzed for the expression of genes specific for growth, gh1 (growth hormone 1), igf1a and igf2a (insulin-like growth factors) [47], and lipid metabolism, fasn (fatty acid synthases) [48]. As described [21], KO juveniles showed reduced igf2a and elevated fasn expression, but no significant changes in gh1 or igf1a compared to the WT (Fig. 3E). Expression of gh1, igf1a and igf2a in the ∆TAD mutant remained unchanged, but fasn was significantly upregulated. ∆TM showed significant downregulation of igf1a expression, whereas expression of gh1 and igf2a remained unaltered, while expression of fasn was upregulated, becoming significantly elevated relative to the KO (Fig. 3E).

Discussion
The critical function of STAT proteins as inducible transcription factors downstream of cytokines is well documented [2,49]. However, there is growing evidence that in addition to such canonical roles, STATs also participate in non-canonical functions [7]. However, the in vivo impact of these non-canonical functions remains poorly understood since most studies are based on KO/LOF mutants that abrogate both canonical and non-canonical functions. Zebrafish Stat5.1 is structurally and functionally conserved with STAT5B, and ablation of either impacted growth, adiposity, and lymphopoiesis [11,21,22]. This study sought to investigate potential non-canonical functions of zebrafish Stat5.1.
Two Stat5.1 mutants were successfully generated using CRISPR-Cas9 genome editing. The first mutant, ∆TAD (p.P701fsX705*), truncated the entire TAD, thereby ablating its ability to act as a transcriptional activator. This is similar to the naturally-occurring splice variants encoding STAT5β proteins that also lack a TAD. These variants are able to bind to the promoter region of responsive genes but not initiate transcription, and potentially hinder the binding of full-length STAT5 proteins to these sites, thereby disrupting canonical functionality [50,51]. The second mutant, ∆TM (p.P701-K704del), deleted four amino acid residues at positions +3 to +6 downstream of the phosphotyrosine (pY) motif that includes three residues within the critical +1 to +5 region responsible for SH2 domain/pY recognition [52][53][54]. Loss of these amino acid residues may impact the specificity of SH2 domain interactions [55] and so perturb functionality. Western blot analysis confirmed the expression of proteins of appropriate sizes and demon-strated that both the ∆TAD and ∆TM mutants retained the ability to be tyrosine-phosphorylated (Supplementary Fig.  1A). The ∆TAD or ∆TM mutants were compared to a recently described Stat5.1 KO/LOF mutant [21] as well as wild-type (WT) fish. If the ∆TAD mutant exhibited a particular phenotype similar to the KO, this was likely due to loss of canonical functionality. Conversely, if it displayed a phenotype similar to the WT, then non-canonical functionality was potentially responsible.
As we have recently described, the Stat5.1 KO mutant exhibited disrupted growth at all developmental stages compared to WT fish with enhanced adiposity observed in adult females along with reduced igf2a and increased fasn [21]. This finding suggests that altered metabolism was responsible, consistent with studies on an alternate Stat5.1 KO mutant [56]. The Stat5.1 ∆TAD mutant also showed significant growth deficiency, increased female adiposity, and altered target gene expression compared to the WT, indicating that loss of canonical functionality was likely responsible for the metabolic perturbation. The zebrafish vizzini line, which harbors a LOF mutation in the zebrafish growth hormone gene paralogue gh1, displayed even more severe growth defects and increased adiposity [57]. This suggests canonical Stat5.1 functionality downstream of growth hormone regulates zebrafish growth and adiposity, and the more severe phenotype in the vizzini line was likely due to redundancy with other downstream pathways.
The Stat5.1 KO mutants further displayed decreased embryonic and adult T lymphopoiesis, but the B and NK cells of these mutants were largely unaffected, as described [21]. The Stat5.1 ∆TAD mutant exhibited a similar significant reduction in embryonic T lymphopoiesis compared to the WT. However, adult lymphopoiesis was impacted less uniformly by the mutants. In adult kidney marrow, T lymphopoiesis was significantly impacted in the ∆TAD mutant compared to the WT, but to a lesser extent than in the KO, whereas in the adult spleen only the KO displayed disrupted T lymphopoiesis. The ∆TAD mutant also showed no alterations in B and NK cell markers. This suggests that canonical functionality predominates in embryonic and some aspects of adult T lymphopoiesis, but that non-canonical functionality might contribute to other facets of adult T lymphopoiesis. LOF mutations of zebrafish il2rg.c, il7r, jak1, or jak3 all impacted embryonic T lymphopoiesis in a similar, if not more extreme manner [31,58,59]. Loss of il2rg.c or jak3 also caused a significant reduction in adult T lymphocyte populations, but B lymphocytes and NK cells were additionally impacted [59,60]. Collectively, this indicates that canonical Stat5.1 functionality downstream of IL-2R family members plays a major role in regulating T lymphopoiesis, but redundant signaling via Stat5.2 and/or other Stat proteins, and parallel signaling pathways such as PI3K and IRS also likely contributes to certain T cell lineages, as well as B and NK cells [61,62].
The mechanism by which non-canonical Stat5.1 functions regulate aspects of adult lymphopoiesis remains to be determined, although precedents exist for non-canonical STAT5 functions impacting mammalian immune cell development. For example, human STAT5B mutations that delete most of the protein-and so likely LOF-are associated with severe T and NK lymphopenia [12,17], but shorter truncations that would still ablate canonical functions result in milder effects, such as reduced CD8 + T cells with normal levels of total T and NK cells [18,19]. Alternatively, suppression of genes mediating Th9 cell differentiation by IL-2-induced pSTAT5 was important for controlling the balance of Th9 versus Th17-like differentiation [63], while IL-7-induced pSTAT5 has been demonstrated to repress the transcription of Igk genes to maintain the proliferation of large pre-B cells and restrict transition to small pre-B cells [64]. Other STATs have also been found to exert their effects via non-canonical functions. For example, uSTAT1 can mediate transcriptional activation of IFN-induced genes independent of IFN stimulation [65][66][67], while uSTAT3 contributes to the expression of various NFκB-dependent genes, as well as oncogenes such as RANTES, IL6, IL8, MET, and MRAS [68,69]. Current therapeutic agents target canonical STAT function, being directed at either upstream signaling components such as JAKs or the SH2 domains of STATs [70]. However, these agents may not be effective against diseases in which non-canonical functions are involved. Elucidation of critical non-canonical functionalities will inform the development of more appropriate therapeutic agents that accurately target specific STAT functions.
The distinct phenotypes of the ∆TM mutant were intriguing. During embryonic lymphopoiesis and for some components of adult lymphopoiesis its phenotype mirrored that of the KO (and ∆TAD) mutant, indicative of LOF. However, in other aspects of adult lymphopoiesis it mirrored the WT (and ∆TAD mutant), suggesting normal function. In contrast, the ∆TM mutant disrupted growth and adiposity to an even greater extent than observed in the KO, and was associated with lower expression of igf1a rather than igf2a, and also higher fasn expression than the KO, revealing possible gain-of-function (GOF) properties. To gain further insights, ∆TM heterozygotes were in-crossed and the progeny were analyzed for rag1 expression prior to genotyping (Supplementary Fig. 1B). This revealed that both homozygous and heterozygous ∆TM mutants showed significantly reduced embryonic T lymphopoiesis compared to the WT, indicating the mutant can act dominantly over the wild-type. Expression analysis was also performed on genes shown to be regulated by Stat5.1 in our recent NGS analysis [21], including the ghrb, srebf1, and cish.a genes, the latter possessing tetrameric STAT5 sites in its promoter able to be bound by Stat5.1, with its expression decreased following Stat5.1 ablation, unlike the related cish.b [71]. The ghrb, srebf1, and cish.a genes, but not cish.b, showed substantial down-regulation in the KO (Supplementary Fig. 1C). Interestingly, while ghrb expression was also significantly decreased in the ∆TM mutant, this was not the case for srebf1 and cish.a, nor for igf2a, another STAT5 target, suggesting that the ∆TM mutant can still activate the transcription of these genes. This contrasts its effects on the fasn gene, also a STAT5 target [72], that displayed even more elevated expression than the KO and igf1a which was reduced compared to WT and KO. This indicates that ∆TM mutants can likely still interact with and regulate at least some direct target genes but differentially, and may also have the capability to interfere dominantly with WT functions on some of these genes. Additionally, it could potentially bind to activated Stat5.2, thereby interfering with its function.

Conclusions
This study explored the relative involvement of canonical versus non-canonical and other functions for zebrafish Stat5.1 in vivo. The results revealed strong impacts of Stat5.1 mutations on embryonic lymphopoiesis, growth, and adiposity that largely appear to be due to loss of canonical functions. However, there was also evidence of potential non-canonical Stat5.1 functions regulating adult T lymphopoiesis and growth, with growth and adiposity also impacted by alternate functions.

Availability of Data and Materials
All data generated or analyzed during this study are included in this article, or available upon request.

Author Contributions
ACW and CL designed the research study. NA and CL performed the research. NA, ACW and CL analyzed the data. NA and ACW wrote the manuscript. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript.

Ethics Approval and Consent to Participate
This study was approved by the Deakin University Animal Ethics Committee under projects 23-2019, 24-2019, 25-2019, 14-2022, 15-2022, and 16-2022. review of this article and has no access to information regarding its peer-review. Full responsibility for the editorial process for this article was delegated to Amedeo Amedei.

Supplementary Material
Supplementary material associated with this article can be found, in the online version, at https://doi.org/10. 31083/j.fbl2808187.