All animal procedures were performed according to protocols approved by the University of California, San Diego Institutional Animal Care and Use Committee and the Institutional Animal Care and Use Committee of Michigan State University and conducted in accordance with the Guide for the Care and Use of Laboratory Animals (48). Mice were maintained on a light/dark cycle of LD [12 h light, 12 h dark, average light intensity ~150-350 lux within the cage]. Based on the newly proposed light reporting method by (49, 50), we determined the relative perception of light by mice using the mouse α-opics equivalent daylight illuminance (EDI, Supplementary Figure 1 ). We calculated that the α-opics for our experimental mice on LD to be melanopsin = 43.2 ± 16.9 lux, rod = 51.6 ± 19.8 lux, s-cone = 0.03 ± 0.01 lux, and m-cone = 57.1 ± 21.7 lux. Lights ON = ZT0 (6:00), lights OFF = ZT12 (18:00), Zeitgeber time (ZT). Vax1^flox/flox (Vax1^{tm1c(KOMP)Mbp}, MGI: 5796178) (51), were crossed with mice heterozygous for the Vip^Cre/wt allele (JAX #010908). Period2::Luciferase (PER2::LUC) mice were purchased from JAX (strain B6.129S6-Per2tm1Jt/J, JAX #006852) and crossed with the offspring of Vax1^flox/flox and Vip^Cre/wt mice. We did not detect any significant differences in mice heterozygous for the Vip^Cre/wt allele as compared to Vax1^flox/flox mice, thus we pooled the Vax1^flox/flox and Vip^Cre/wt control groups into one control group, referred to as Ctrl. Genotyping primer sequences were as follows: Vax1-wtF: CCAGTAAGAGCCCCTTTGGG, Vax1-floxF: GCCGGAACCGAAGTTCCTA; Vax1-R: CGGATAGACCCCTTGGCATC; CreF: GCATTACCGGTCGTAGCAACGAGTG, CreR: GAACGCTAGAGCCTGTTTTGCACGTTC. Mice were kept on a C57BL/6J background and were screened for germline recombination. Mice with germline recombination were excluded from the studies. VIP-tdTomato mice were generated by crossing Vip^Cre/wt mice with Rosa-tdTomato reporter mice (JAX# 007914). Mice were euthanized by cervical dislocation followed by decapitation.

Female and male mice aged 8-12 weeks at the start of the experiment were single-housed in cages containing metal running wheels and wheel revolutions were monitored using magnetic sensors. All cages were contained in a light-tight cabinet with programmable lighting conditions and rooms were monitored for temperature and humidity. Sylvania T8 32-Watt 4100K fluorescent bulbs (F032/841/ECO) were used to provide light to the cabinets. Food and water were available ad libitum during the entire experiment. After 1-week acclimation to the polypropylene cages (17.8 × 25.4 × 15.2 cm or 33.2 × 15 × 13.2 cm) containing a metal running wheel (11.4 cm diameter or 11 cm diameter, respectively), locomotor activity rhythms were monitored with a VitalView data collection system (Version 4.2, Minimitter, Bend OR) that integrated into 6 minute bins the number of magnetic switch closures triggered by half wheel rotations or full wheel rotations, respectively. Running wheel activity was initially monitored for 2 weeks in a standard 12 h light/12 h dark cycle (LD). Subsequently, mice were monitored for 4 weeks in constant darkness (DD), with wheel running data analyzed from weeks 2-4 (14 days) in DD. Cage changes were scheduled at 3-week intervals. The light intensity varied between 268-369 lux inside the mouse cages. Wheel running activity was analyzed using ClockLab Analysis (ActiMetrics) by an experimenter blind to the experimental group. The circadian period was analyzed by constructing a least-squares regression line through a minimum of 13 daily activity onsets. Daily onset and offset of activity, defined as a period of 5 h of activity following 5 h of inactivity (onset) or a period of 5 h of inactivity following 5 h of activity (offset), were used to calculate the length of the active phase (alpha). Chi² periodograms were generated for periods from 0 to 36 h, with significance set at 0.001. Activity profiles were generated for weeks 2-4 in DD using the estimated chi² periodogram tau for the same time period. Total daily counts for mice on wheels with 2 sensors were calculated over 24 h, during both LD and DD.

Female mice aged 9-13 weeks were single-housed for two weeks on LD, then placed in a transparent 7 by 24 inch cylinder filled 2/3 full with 20-21°C water. Swim tests were completed during the mice active phase and started at ZT13. Mice were recorded in dim red light (5 lux, α-opics: melanopsin = 0.03, rod = 0.20, s-cone = 0, m-cone = 0.54, Supplementary Figures 1C, F ) for 6 minutes before they were removed from the water, dried with a cloth, and returned to their cages for one hour before euthanasia and blood collection from the abdominal aorta. Videos were scored by an observer blinded to the experimental group using the sampling method where the mouse is determined to be either floating or swimming every 30 seconds for 5 minutes, with the first minute of each video going unscored and serving as an adjustment period. The percent of intervals where the mouse was observed to be floating is reported (52).

For the fertility assessment, virgin 8- to 12-week-old male and female Ctrl, and Vax1^Vip mice were housed with opposite-sex Ctrl mice (53, 54). The number of litters and the number of pups per litter were recorded over a period of 4 months, as described previously (54). Estrous cyclicity was monitored by vaginal lavage with 20 µl H₂O daily between ZT3 and ZT5 for 16-18 days. The lavage solution was dried on a slide and stained with 0.5% methylene blue. Cytology was visually examined and scored. Ovary and uterus weights were collected after euthanasia in diestrus. Following euthanasia in males, testes and epididymis were collected and weighed. Sperm was collected from the epididymis of male mice in M2 media (Sigma #M7167). The epididymis was cut in half and sperm were expelled by gently pressing down on the epididymis and then left in M2 media at room temperature for 15 min. The numbers of total and motile sperm were counted from a 1:10 dilution of the M2 media containing sperm by using a hemocytometer. The second epididymis was cut into small pieces and left 15 minutes at room temperature in M2 media. The solution was homogenized frequently to help liberate the sperm. The solution was filtered using a cell streamer (70 µm, Falcon #352350) and sperm were diluted 1:10 with MQ before counting total number of sperm heads.

Pubertal onset was established by visual inspection of preputial separation (PPS) in males and vaginal opening (VO) in females, as described previously (54). Body weight was recorded daily until pubertal onset was observed.

Tissues were collected between ZT3 and ZT5 from adult male and proestrus female mice on LD light cycle and fixed overnight at 4°C in 60% ethanol, 10% formaldehyde, and 10% glacial acetic acid. Tissues were washed in 70% ethanol and embedded in paraffin. Single immunohistochemistry on 10 µm coronal brain sections embedded in paraffin was performed as previously described (24). The primary antibody was rabbit anti-VIP (Immunostar #20077, 1:1000, RRID : AB_572270). Sections were incubated in 1:300 secondary anti-rabbit IgG (Vector Laboratories, #BA-1000). Secondary antibodies were purchased from Vector labs, and colorimetric VIP (purple staining) and DAB (brown staining) assays (Vector laboratories) revealed the primary antibodies.

Vip^Cre/wt:Rosa-tdTomato^+/- mice (n=4; 2 female and 2 male) were sacrificed between ZT4-6 at 6 weeks of age and brains immersed in 4% PFA overnight at 4°C. Brains were transferred to 30% sucrose until sectioned 40 μm thick with a cryostat. Sections were stored in cryoprotectant at -80°C. Prior to staining, sections were washed overnight in PBS at 4°C. Sections underwent antigen retrieval for 20 minutes in citrate buffer, followed by a wash and blocking for 30 minutes at room temperature using an Avidin/Biotin blocking kit (Vector Labs) with 5% normal goat serum. Sections were briefly washed and then stained following the protocol for the Mouse on Mouse Basic Immunodetection kit (Vector Labs) using a mouse anti-VAX1 (1:100; Origene RRID : AB_2941013). Slices underwent Vectastain ABC kit (Vector Labs) followed by TSA treatment (Akoya Biosciences) for 10 minutes and finally streptavidin-conjugated secondary (1:200) for 30 minutes. Slices were mounted, air-dried, and coverslipped with Prolong Gold with DAPI (ThermoFisher). Slices were imaged on a Nikon Eclipse Ti2-E using a Lumencor SpectraX LED and acquired using a DS-Qi2 CMOS camera. One SCN section per animal was analyzed using QuPath (55). All image manipulations were applied homogenously to the entire image.

To examine Vip, Avp, Nms, and Vax1 mRNA when adult male and female hormones are most comparable, brains were collected at ZT4-8 in young mice, adult males, and diestrus females. To examine Avp, Vip, and Bmal1 mRNA around the time of the LH surge in Ctrl and Vax1^Vip mice, brains were collected at ZT13-16 in proestrus females. Multiplex in situ hybridization detection of mouse (Mus musculus) mRNAs was performed with RNAscope® LS Multiplex Fluorescent Reagent Kit (Advanced Cell Diagnostics, cat no. 322800) for 3-plex assay in addition to RNAscope® LS 4-Plex Ancillary Kit (Advanced Cell Diagnostics, cat no. 322830) for 4-plex assay following vendor’s standard protocol for FFPE tissue sections with minor modifications. RNAscope® assays were performed on a Leica Bond autostainer as described (56) with the following probes: RNAscope® 2.5 LS Probe – Mm-Arntl (also known as Bmal1) [aryl hydrocarbon receptor nuclear translocator-like (Arntl) transcript variant 1 mRNA, cat no. 438748-C1] or RNAscope® 2.5 LS Probe - Mm-Vax1 mRNA – [musculus ventral anterior homeobox 1 (Vax1), cat no. 805108-C1]; RNAscope® 2.5 LS Probe – Mm-Avp-C2 [arginine vasopressin (Avp) mRNA, cat no. 401398-C2]; RNAscope® 2.5 LS Probe – Mm-Vip-C3 [vasoactive intestinal polypeptide (Vip) mRNA, cat no. 415968-C3]; and RNAscope® 2.5 LS Probe – Mm-Nms-C4 [neuromedin S (Nms) transcript variant 1, cat no. 472338-C4]. Stock Mm-Nms-C4 probe was diluted at 1:50 in a pre-diluted C1 probe as recommended by the vendor, whereas stock Mm-Avp-C2 and Mm-Vip-C3 were further diluted to 1:100 in appropriate pre-diluted C1 probe due to saturating signal in the pilot experiment. Tissue slides were counterstained with DAPI and scanned with an Aperio Versa imaging system with 20X objective with customized narrow-width band excitation and emission filter cubes as described (56). The Aperio Cellular IF Algorithm (Leica Biosystems, No: 23CIFWL) was used for automated cell enumeration and segmentation based on nuclear DAPI staining. Cells were classified based on the expression levels of one or more mRNAs. In images taken at P2, Vax1 staining was oversaturated, so they were re-imaged for proper mRNA visualization. Representative images at P2 were displayed with increased contrast, applied to all channels to compare with P10 and P60 in Figure 1 .

Hormonal challenges were done using kisspeptin and GnRH intraperitoneal (i.p.) injections at ZT3-4. Kiss-10 (catalog #42-431, Batch 7A, Fisher Science, 30 nmoles/mouse) was injected into males or diestrus females and blood was collected from the mouse from the tail vein before (time 0) and after i.p. injection at time points 5, 10, 15, 30, and 45 minutes. Tail blood was collected before (time 0) and 10 minutes after i.p. injection of GnRH (Millipore Sigma, catalog #L7134, 1 μg/kg dose). For all other serum hormone analyses the mice were killed by cervical dislocation, and blood was collected from the abdominal aorta between ZT3 and ZT6. Blood was allowed to clot for 1 hour at room temperature, then centrifuged (room temperature, 15 minutes, 2,600× g).

Serum was collected and stored at -80°C before analysis for estradiol at the Center for Research in Reproduction, Ligand Assay, and Analysis Core, University of Virginia (Charlottesville), by Luminex analysis for LH and FSH on MILLIPLEX MAP Mouse Pituitary Magnetic Bead Panel (Millipore Sigma #MPTMAG-49k) or a competitive enzyme-linked immunosorbent assay (ELISA) kit (EIACORT, ThermoFisher) for corticosterone. Coefficients of variance (CVs) were based on the variance of samples in the standard curve run in duplicate. Reportable range: estradiol: 3–300 pg/ml, CV = <20%; LH: lower detection limit: 5.6 pg/ml, CV < 15%; FSH: lower detection limit: 25.3 pg/ml, CV < 15%; corticosterone: lower detection limit: 0.87 μg/dl, CV = <20%. Samples were run in singlets.

For circadian rhythm organotypic explant studies, tissues from mice expressing the PER2::LUC circadian reporter were collected and analyzed as previously described (57). Male and proestrus and diestrus PER2::LUC females were placed on LD and euthanized at ZT3-4 via isoflurane inhalation and cervical dislocation. The brain was removed immediately and placed in an ice-cold, CO₂-saturated Hank’s Balanced Salt Solution (HBSS) for approximately 1 hour. Using a Vibratome (Leica), coronal brain sections of 300 μm were collected and the SCN was dissected from the slices in ~2x2 mm squares and placed on a 30 mm Millicell membrane (Millipore-Sigma) in a 35 mm cell culture plate containing 1 mL Neurobasal-A Medium (Gibco) with 1% Glutamax (Gibco), B27 supplement (2%; 12349-015, Gibco), and 1 mM luciferin (BD Biosciences). The lid was sealed to the plate using vacuum grease to ensure an air-tight seal. Plated tissues were loaded into a LumiCycle luminometer (Actimetrics) inside a 35˚C non-humidified incubator at ZT6-6.5, and recordings were started. The bioluminescence was counted for 70 seconds every 10 minutes for 6 days (day 1 – day 7 of recording time). PER2::LUC rhythm data were analyzed using LumiCycle Analysis software (Actimetrics) by an experimenter blind to the experimental group. Data were detrended by subtraction of the 24 h running average, smoothed with a 2 h running average, and fitted to a damped sine wave (LM Fit, damped). The period was defined as the time in hours between the peaks of the fitted curve. Amplitude was defined as the value of the second peak and phase was defined as the time of the first peak. Data from proestrus and diestrus female SCN recordings were pooled as no significant differences in PER2::LUC period or amplitude were found.

NIH3T3 (American Type Culture Collection) and COS-1 (American Type Culture Collection) cells were cultured in DMEM (Mediatech), containing 10% fetal bovine serum (Gemini Bio), and 1x penicillin-streptomycin (Life Technologies/Invitrogen) in a humidified 5% CO₂ incubator at 37˚C. For luciferase assays, NIH3T3 cells were seeded into 24-well plates (Nunc) at 30,000 cells per well. For electrophoretic mobility shift assays (EMSA) COS-1 cells were plated at 1.5 million cells/10 cm dish. Transient transfections for luciferase assays were performed using PolyJet™ (SignaGen Laboratories, Rockville, MD), whereas Fugene was used for plasmid overexpression for EMSA, following the manufacturer’s recommendations. Transfection of cells was performed 48 h after the cells were plated. COS-1 cells were transfected with Vax1/DKK-Flag or CMV6/DKK-Flag overexpression plasmids (20 ng/well, Origene Technologies, Rockville, MD) and harvested at sub-confluency 48–56 h after transient transfections in 10 cm dishes (Nunc). Transient transfections for luciferase assays were done following the manufacturer’s recommendations. NIH3T3 cells for luciferase assays were co-transfected with 150 ng/well of Bmal1-luciferase or Per2-luciferase reporters, 100 ng/well thymidine kinase-β-galactosidase reporter plasmid, which served as an internal control (54), as well as mouse Vax1/pCMV6 overexpression plasmid (20 ng/well, Origene Technologies, Rockville, MD), or its empty vector control (pCMV6). To generate the Bmal1-luciferase plasmid the Bmal1 sequence between -966 bp to +140 bp from the Bmal1 transcriptional start site was excised from the pABpuro-BluF plasmid (Addgene, Plasmid #46824) with PCR primers (F: gggctacaacagaacaactaac, R: taaacaggcacctccgt). The PCR product was inserted into the pGL3-basic backbone between the Mlu-HF and XhoI sites using the Quick Ligation Kit (New England Biolabs). Site directed mutagenesis of the homeodomain binding sites (ATTA and ATTA-like) in the mouse Bmal1-luciferase plasmid was performed using the NEB Q5 Site-Directed Mutagenesis Protocol (New England Biolabs Inc.), following manufacturer’s instructions. Primers for NEB Q5 site-directed mutagenesis were designed using the NEB Base Changer ( Table 1 ). To equalize the amount of DNA transfected into cells, we systematically equalized plasmid concentrations by adding the corresponding inactive plasmid backbone. Cells were harvested 24 h after transfection in lysis buffer [100 mM potassium phosphate (pH 7.8) and 0.2% Triton X-100]. Luciferase values were normalized to β-galactosidase values to control for transfection efficiency. Values were further normalized by expression as fold change compared to the pGL3 control plasmid, as indicated in the figure legends. Data represent the mean ± SEM of at least three independent experiments done in duplicate and triplicate.

COS-1 cells were scraped in hypotonic buffer (20 mM Tris-HCl, pH 7.4, 10 mM NaCl, 1 mM MgCl₂, 10 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 1x protease inhibitor cocktail; Sigma-Aldrich) and left on ice to swell. Cells were lysed and nuclei were collected by centrifugation (4°C, 1700 g, 4 minutes). Nuclear proteins were extracted on ice for 30 minutes in hypertonic buffer [20 mM HEPES, pH 7.9, 20% glycerol, 420 mM KCl, 2 mM MgCl₂, 10 mM NaF, 0.1 mM EDTA, 0.1 mM EGTA, 1x protease inhibitor cocktail (Sigma-Aldrich), and 1 mM phenylmethylsulfonylfluoride]. Debris was eliminated by centrifugation (4°C, 20,000 g, 10 min), and the supernatant was snap-frozen and stored at -80°C.

Oligonucleotide probes are listed in Table 1 . All synthetic oligonucleotides were made by IDT (San Diego, CA). Annealed double-stranded oligonucleotides (1 pmol/µl) were end-labeled with T4 Polynucleotide Kinase (New England Biolabs, Ipswich, MA) and [γ³²P]ATP (7000 Ci/mmol; MP Biomedicals, Solon, OH). Probes were purified using Micro Bio-Spin 6 Chromatography Columns (Bio-Rad). Binding reactions contained 2 µg nuclear protein and 1 fmol of labeled probe in 10 mM HEPES (pH 7.9), 25 mM KCl, 2.5 mM MgCl2, 1% glycerol, 0.1% Nonidet P-40, 0.25 mM EDTA, 0.25% BSA, 1 mM dithiothreitol, and 350 ng poly(dI-dC). For super-shift experiments, 2 µg mouse anti-DKK (Flag antibody, Origene #TA50011) or 2 µg of normal mouse IgG (Santa Cruz Biotechnology, #sc2025) were added to the reaction. Samples were incubated for 20 minutes at room temperature before loading on a 5% non-denaturing polyacrylamide gel in 0.25x Tris-borate EDTA buffer. Gels were run for 2 h at 200 V, dried under vacuum, and exposed to film for 2-5 d at room temperature.

Statistical analyses were performed with GraphPad Prism 8, using Student’s t-test, one-way ANOVA, or two-way ANOVA, followed by post hoc analysis by Tukey or Bonferroni as indicated in figure legends, with p < 0.05 to indicate significance. All data were analyzed as independent measures except for wheel-running activity, which was analyzed via a two-way repeated-measures ANOVA. PER2::LUC timing of first peak phase relationships was analyzed in R via a Circular Analysis of Variance High Concentration F-Test, with a corrected confidence level of p < 0.01667 to account for family-wise error.

Vax1 is highly expressed in the developing mouse SCN and becomes refined to the hypothalamus, primarily in the SCN in the early postnatal period (25, 58). Although conditional deletion of Vax1 in late neuronal development using the Synapsin^Cre allele reduced VIP expression in the adult SCN (25), it remains unknown if all VIP expressing neurons co-express VAX1 and how postnatal deletion of Vax1 in VIP neurons impacts VIP expression. Because VAX1 is highly expressed in the developing SCN, we first asked how VAX1 expression changed after birth and into adulthood in males and females. To answer this, we performed multiplex RNAscope® assay at postnatal day 2 (P2), P10, and P60 (adult) at ZT5 in males and females. We found that all SCN Vip-expressing cells at P2 co-express Vax1 [ Figures 1A, B, G ; male (n = 2) 99.4 ± 0.6%, female (n = 2) 100% ± 0%], a pattern maintained at P10 [ Figures 1C, D, I ; male (n = 4) 99.98 ± 0.021%, female (n = 4) 99.91 ± 0.09%], and P60 [ Figures 1E, F, K ; male (n = 3) 98.80 ± 1.08%, female (n = 4) 100 ± 0%]. In addition to the high co-expression with Vip, Vax1 is highly expressed throughout the SCN at P2, P10, and P60 ( Figures 1A–F ), where both Avp and Nms-expressing cells also exhibited full overlap with Vax1 in both sexes ( Figures 1G, I, K ). Interestingly, we found a sex difference in the number of cells expressing Avp at P60, where females had fewer Avp+ cells compared to males [ Figure 1L ; n = 7, t(5) = 5.671, p = 0.0024], a difference that was not present prior to puberty (P10). Although we did not see a significant sex difference in the number of Vip neurons at any age, the concentration of Vip, as evaluated through Vip probe signal intensity, was significantly lower in females than males at P10 [t(3) = 6.01, p = 0.037, not shown] and trended lower at P60 [t(5) = 2.47, p = 0.16, not shown]. To determine if this modest sex difference in Vip concentration translated to a sex difference in peptide levels, we performed IHC in adult male and female brains. Adult female mice had a significant reduction in the intensity of VIP peptide in the SCN [ Figures 2A, B , t(11) = 3.874, p = 0.0026] as compared to males.

As Vax1 is ~100% co-expressed with Vip from P2 until P60 in males and females ( Figure 1 ), we next generated a conditional knockout mouse to determine how the loss of VAX1 in VIP neurons would impact SCN function. Using our RNAscope^® ISH data from Figure 1 , we first visually inspected all stained sections for potential Vax1 expression in non-SCN Vip cells. The olfactory bulb is the only additional brain area that expresses Vax1 which is also targeted by the Vip^Cre allele ( Figure 2C ). In the scenario, the Vip^Cre allele does target some Vax1 expressing cells in the olfactory bulb, a change in reproductive behavior could impact fertility data because olfaction is required for normal male reproductive behavior. To validate that the Vip^Cre allele targeted VAX1 expressing neurons of the SCN, we generated Vip^Cre:RosaTdTomato (Vip^Cre:Td) mice allowing the identification of all neurons that are targeted by the Vip^Cre allele ( Figures 2D, E ). Using dual IHC, we quantified the colocalization of VIP- and VAX1-expressing neurons in the SCN. Using tdTomato as a marker for VIP neurons in sections from Vip^Cre:Td animals, we found 38 ± 7% of SCN tdTomato+ neurons colocalized with VAX1 (n=4, 2 per sex, 36-101 neurons per animal). An average of 23 ± 5% of VAX1 neurons colocalized with tdTomato (n=4, 65-236 neurons per animal). This discrepancy between the RNAscope and dual IHC has numerous potential explanations, as detailed in the discussion.

To determine the role of VAX1 in VIP neuron circadian output, we evaluated the wheel running behavior of Ctrl and Vax1^flox/flox:Vip^Cre (Vax1^Vip) mice in LD and constant darkness (DD). Wheel running patterns in LD of Ctrl and Vax1^Vip females and males were comparable ( Figures 3A–I , LD). Light is a strong entraining signal of the SCN, and activity rhythms in LD can mask weakened SCN function (59). Following the initial LD period, mice were placed in DD for 28d to assess the endogenous free-running period. Both male and female Vax1^Vip mice showed a significantly shortened free-running period (Tau) compared to Ctrls ( Figure 3E , two-way ANOVA, male p = 0.0004, female p = <0.0001) with no differences in Chi² amplitude (Qp). There were no changes in the number of wheel revolutions per day or activity duration (alpha, Figures 3F–I ). To determine if the shortening of the free-running period during DD resulted from a change in the endogenous SCN circadian period, we generated triple transgenic mice crossing Vax1^Vip mice with the PER2::LUC reporter mouse (60). In agreement with the significantly shortened behavioral period of Vax1^Vip males and females on DD, we found that the SCN of Vax1^Vip:PER2::LUC mice showed a significant shortening in period as compared to Ctrl males [ Figure 4A , t(27) = 2.936, p = 0.0067] and females [ Figure 4B , t(24) = 2.125, p = 0.0440]. No differences were found in the amplitude or phase relationships of PER2::LUC in the SCN of Vax1^Vip males or females as compared to Ctrl, indicating that the rhythms in the SCN are not misaligned or significantly weakened ( Figures 4C–F ). Together these data show that loss of VAX1 in VIP neurons shortens SCN circadian output in both males and females.

A shortened SCN period can be driven by both changes in SCN peptide expression and changes in molecular clock gene expression. We have previously shown that VAX1 promotes Per2-luciferase plasmid expression using transient transfection assays (25). However, we have not yet demonstrated whether this action is direct or indirect. To assess if VAX1 directly binds with our top candidate ATTA site of the mouse Per2 DNA regulatory region (25), we used EMSA. We found that VAX1 directly binds the ATTA site at +1774/1770bp from the transcriptional start site of the mouse Per2 gene ( Figure 5A , Super shift is indicated by *). In addition to ATTA sites in the Per2 regulatory region, the Bmal1 regulatory region also contains numerous ATTA sites. Using transient transfections, we found that VAX1 promotes Bmal1-luciferase expression ( Figure 5B ). Site-directed mutagenesis of ATTA-like sites in the Bmal1 regulatory region ( Table 1 ) showed a modest increase in fold-change of VAX1-driven Bmal1-luciferase expression in transfected cells ( Figure 5B ). Interestingly, VAX1 can directly bind to all the identified ATTA sites tested by EMSA (EMSA, Figure 5C , supershift indicated by *). This identifies for the first time that VAX1 can directly bind to the regulatory regions of Per2 and Bmal1 and provides a mechanism by which changes in VAX1 expression can directly impact molecular clock function. Next, to determine if loss of VAX1 in VIP neurons would significantly impact Bmal1 expression in Vip neurons, we performed RNAscope^® for Bmal1, Vip, and Avp in the adult SCN of Ctrl and Vax1^Vip females ( Figure 6 ). Note these experiments were completed in proestrus at ZT16, with the goal of having the most hormonally challenging environment in the female body present at the time of sample collection. We found that in Ctrl and Vax1^Vip females almost 100% of Vip and Avp neurons co-expressed Bmal1 ( Figures 6A, B ). Despite a trend in a reduction in Vip in the SCN of Vax1^Vip females ( Figure 6C ), as well as a trend in the reduction in cells co-expressing Vip and Bmal1 ( Figure 6D ), no significant difference in any of the studied transcripts, or colocalization of transcripts were identified ( Figures 6C, D ).

The SCN provides daily neuronal and hormonal signals aligning circadian timekeeping in peripheral reproductive tissues allowing coordination between hormone release and increased tissue sensitivity improving reproductive function (25, 27, 35, 57). To determine if Vax1^Vip mice have misaligned circadian phase of their reproductive tissues, we recorded PER2::LUC expression in the pituitary, ovary (female), uterus (female), and epididymis (male) of triple transgenic mice. No tissues were found to have significant differences in period ( Table 2 ), amplitude ( Table 2 ), or time of first PER2::LUC peak ( Table 2 , phase). These data indicate that the weakened SCN output of Vax1^Vip mice does not significantly impact circadian timekeeping in the studied peripheral tissues but does not preclude disruptions to fertility. To determine if reproductive function is impacted in Vax1^Vip mice, we first evaluated pubertal onset. At pubertal onset, body weight was comparable between Ctrls and Vax1^Vip in both sexes ( Table 3 ). Male pubertal onset, as evaluated by preputial separation (PPS) was slightly delayed [ Figure 7A , t(18) = 2.211, p = 0.040], whereas female pubertal onset, as assessed through vaginal opening (VO) and first estrous, were comparable between Ctrls and Vax1^Vip females ( Table 3 ). There was no impact on male reproductive function ( Table 3 ) apart from significantly increased total sperm count [ Figure 7B , t(12) = 3.101, p = 0.009]. Vax1^Vip males had normal testis size and percent motile sperm ( Table 3 ). The increase in total sperm pool was not associated with changes in basal LH and FSH levels in males ( Table 3 ). Both Vax1^Vip males and females were comparable to Ctrls for the number of litters generated in 90 days, days to first litter, and litter sizes ( Table 3 ). Despite the normal fertility in females, Vax1^Vip females had a significant lengthening of the estrous cycle [ Figure 7C , t(19) = 2.307, p = 0.033] with a similar amount of time spent in each cycle stage as compared to Ctrls [Two-way ANOVA, F (1, 10) = 2.500, P=0.1449]. This lengthening in estrous cycles was associated with a reduction in FSH, estrogen, and ovarian weight in Vax1^Vip females ( Figures 7D–F ), but did not impact basal LH or uterine weight in diestrus ( Table 3 ).

VIP neurons from the SCN project directly to GnRH neurons and indirectly through AVP to kisspeptin neurons. The anterior pituitary releases LH into the circulation upon GnRH release at the median eminence, allowing an indirect approach to study GnRH neuron function. To determine if the reduction in VIP in the Vax1^Vip SCN impacted GnRH neuron response to kisspeptin, we performed hormone challenges in mice. We first confirmed that the pituitary responded to GnRH by increasing LH release through an i.p. injection of GnRH. As expected, the fold change in LH in response to a GnRH challenge was comparable between Ctrl (9.07 ± 1.97, n = 4) and Vax1^Vip males [12.09 ± 2.43, n = 6, t(8) = 0.885, p= 0.401], and between Ctrl (10.50 ± 3.37, n=8), and Vax1^Vip females [6.95 ± 3.84, n = 4, t(10) = 0.641, p= 0.535]. To assess if the GnRH neuron response to kisspeptin was impacted in Vax1^Vip males and females, we next performed an i.p. kisspeptin challenge. There were no differences between LH release in Vax1^Vip (5798 pg/mL ± 583, n = 4) males and Ctrls [5325 pg/mL ± 670, n = 5, t(7) = 1.11, p = 0.303]. In contrast, Vax1^Vip females had an increased release of LH in response to kisspeptin at 5 and 10 minutes as compared to Ctrls ( Figure 7G , mixed-effects analysis, 5 minutes p = 0.0102, 10 minutes p = 0.0257), as well as an overall increase in LH release [ Figure 7H , t(7) = 2.560, p = 0.0376]. Such alteration in the neuroendocrine network regulating LH release would be expected to impact female estrous cyclicity, which relies on precisely timed hormone release and sex-steroid-feedback.

AVP is expressed outside the SCN and is highly expressed in the paraventricular nucleus (PVN), a direct target of SCN neurons. The PVN is a well-established relay site playing a significant role in several autonomic functions, including stress (61–63). To determine how Avp levels were impacted in the hypothalamus of Vax1^Vip females, we analyzed Avp using RNAscope^® assay. Interestingly, we found that Vax1^Vip females, which have comparable Avp mRNA in the SCN to Ctrl ( Figures 6C, D ), display a significant increase in Avp transcript and cell numbers in the PVN ( Figures 8A–C ). This increase of Avp in the PVN of Vax1^Vip females provides a potential link to the activation of the stress axis. In agreement with this, basal corticosterone levels were overall increased at ZT3 in Vax1^Vip mice [ Figure 8D , two-way ANOVA, F(1,12) = 5.868, p = 0.032]. Finally, to determine if these known risk factors for depression would reflect an increase in depressive-like behavior in Vax1^Vip mice, we tested males and females in the Porsolt forced swim test. We found that Vax1^Vip females (metestrus, ZT13) exhibited increased depressive-like behaviors as shown by an increase in the percentage of time floating ( Figure 8E ) as compared to Ctrl females [t(17) = 2.121, p = 0.0489], while Vax1^Vip males were comparable to Ctrls [t(11) = 0.5889, p = 0.5678]. The increased time floating in the Porsolt forced swim tests of the Vax^Vip females correlated with increased circulating corticosterone 1 h after the swim test [ Figure 8F , t(11) = 3.595, p = 0.0042].

The SCN retains a high expression of numerous homeodomain transcription factors after development (26, 46, 64–66), including Vax1. Vax1 is critical in brain and neuronal development (25, 46, 47, 53, 67) and highly expressed in the developing mouse brain before it becomes refined to the adult hypothalamus, primarily in the SCN, during the early postnatal period (25, 58). We have previously shown that conditional deletion of Vax1 in late neuronal development using the Synapsin^Cre allele reduced VIP expression in the adult SCN (25), but this previous study did not address how many VIP neurons co-expressed Vax1. Here we find that ~100% of Vip-expressing neurons co-express Vax1 at P2 and maintain a close to 100% co-expression at P10 and P60. In addition, we show that Vax1 is also co-expressed by ~100% of Nms and Avp neurons at P2, P10, and P60. This shows for the first time that Vax1 is highly expressed in three primary SCN neuron populations from the early postnatal period into adulthood, suggesting a role of VAX1 in regulating the function of these neuronal populations. The close to 100% co-expression of Vax1 and Vip strengthens the value of our conditional knock-out model as a novel tool to specifically study SCN VIP neurons. This high level of overlap, coupled with the primary restriction of Vax1 expression within VIP neurons in adulthood to the SCN (25, 58), allows for targeted impairment of VIP neurons within the SCN alone. This approach allows us to build upon previous work that identified the importance of VIP in circadian and reproductive function using VIP knockout mice (36), while avoiding some pitfalls of other methods of selectively targeting VIP neurons within the SCN, such as damage to other neurons or brain nuclei via surgical methods due to the ventral location of the SCN (9, 68). One concern with conditional knockout mice is that off-target recombination may impact other brain or peripheral functions or have a negative impact on development. In our model, we found Vip^Cre-mediated recombination outside of the SCN was restricted to the olfactory bulb, which also expresses Vax1. However, as a functioning olfactory bulb is pivotal to male mating (69) and our male mice bred normally, it is likely that the olfactory bulb is not significantly affected in this model. We also generated Vip^Cre : Td mice to identify neurons that were targeted by the Vip^Cre allele prior to tissue collection to validate that the Vip^Cre allele specifically targeted VAX1 expressing neurons of the SCN. We confirmed that SCN tdTomato+ neurons colocalized with VAX1, and VAX1 neurons colocalized with tdTomato, although these percentages were lower than predicted by RNAScope^® ISH assay. The differences in co-localization between mRNA and protein results could reflect different sensitivity limits between the visualization approaches, or alternatively may provide evidence that not all Vax1 mRNA is being translated into protein. Another possibility is that VAX1 expression could be circadian at the mRNA and/or protein level, leading to differences in expression that are time-of-day dependent, although future studies will be needed to determine this.

As a transcription factor that binds to ATTA and ATTA-like sites, a common sequence in the DNA, VAX1 has a high number of genes it can potentially regulate. Here, we build upon our previous work that determined the ability of VAX1 to promote Per2-luciferase expression (25) by providing evidence that this occurs via direct binding of VAX1 to regulatory regions of Per2. In addition to Per2, another core component of the molecular clock, Bmal1 also contains numerous ATTA-like sites in its regulatory region. We found here that VAX1 can directly bind to all the identified Bmal1 ATTA sites tested by EMSA, in addition to promoting Bmal1-luciferase expression. This work provides a mechanism by which changes in VAX1 expression can directly impact molecular clock function by VAX1 binding to the regulatory regions of Per2 and Bmal1. Excitingly, Vax1^Vip mice presented with a shortening in SCN PER2::LUC period in ex vivo recordings, as well a shortened free-running period, together supporting a role of VAX1 as a novel regulator of both molecular clock expression and function. Future work will be required to determine if a loss of VAX1 in VIP neurons causes a circadian phase shift in VIP neurons, and/or if a loss of VAX1 leads to a reduction in molecular clock transcript expression, which could impact clock-controlled gene expression and phase.

VIP neurons within the SCN are an important coordinator of the circadian timekeeping system (37). Given this, it is no surprise that the Vax1^Vip mouse model demonstrates altered SCN output, as indicated by the shortened Vax1^Vip free-running period in both sexes. This finding is consistent with work done by others indicating that a decrease in VIP within the SCN results in a shortened free-running period (36). However, VIP rarely acts alone in the regulation of circadian behaviors, and other peptides and components of the molecular clock exert strong influences on locomotor period (70). A shortened free-running period can be caused by an increase in SCN AVP (71–73) or reductions in Bmal1 expression (74). Others have found that Bmal1 is expressed rhythmically in Avp- and Vip-expressing neurons (75) and that deletion of Bmal1 from Avp-expressing neurons can lengthen free-running periods in mice (76). As Vax1^Vip females do not show significant changes in Avp in the SCN, and only trended towards decreased Vip expression, in addition to a non-significant reduction in neurons co-expressing Bmal1 and Vip, the cause of the shortened period in Vax1^Vip mice remains unknown. Future work will aim at determining if these trending reductions in Vip, combined with a trend in reduced Bmal1 expression in Vip neurons together might contribute to the shortened SCN period of Vax1^Vip mice.

Sex differences are well-documented in SCN morphology and cellular function, previously reviewed in several studies (77, 78). There is strong evidence that VIP is sexually dimorphic, with increased VIP expression in human males (79, 80), as well as increased Vip transcript in male nocturnal laboratory rats (81) and diurnal Nile Grass rats (82). Our data support and extend these findings, where Ctrl male mice also exhibit increased Vip transcript and VIP compared to females. A potential mechanism guiding this sex difference could rely on the influence of the gonadal hormones estradiol and testosterone, which modulate VIP expression in the SCN (81–84). Our data, and others, suggest that sex differences in both Vip transcript and peptide levels (78) occur post-puberty. Recent work has demonstrated detailed spatial patterning of the onset and development of Avp and Vip transcription (85). Our data support the conclusion that SCN Vip neuron development is not complete at P10; however, we find decreased Vip transcript and protein in the adult female SCN compared to males, while increased VIP-TdT+ cell numbers were found in a lateral cluster of the SCN (85). There are several potential rationales for these differences, including delays between transcription onset and Vip^Cre-driven TdTomato expression that may lead to different results from our mRNA transcript measure. Nevertheless, these neuropeptide sex differences highlight the need for continued investigation in both males and females to further our understanding of how SCN function drives circadian output in both sexes and the role of sex steroids therein.

Circadian timekeeping is essential for coordinating hormone release and increased tissue sensitivity within reproductive tissues (25, 27, 35). The SCN modulates the timing of hormone release through direct and indirect projections to GnRH neurons, which in turn regulate the release of LH and FSH (86–88). These hormones are required for reproductive health through the production of testosterone and spermiogenesis in males and ovulation, embryo implantation, and follicular health in females (89–91). Though more frequently studied in females, current studies suggest that severe circadian disruption, through changes either in light exposure or via direct disruption to the SCN, may have a mild influence on male fertility (92). Interestingly, we found that Vax1^Vip males had delayed pubertal onset. Given the importance of both SCN Avp and Vip in LH release (32, 93, 94), a required hormone for pubertal onset (95), it will be of interest for future studies to examine SCN neuropeptide expression in Vax1^Vip males undergoing puberty. Interestingly, aside from delayed pubertal onset, there were no significant deficits in Vax1^Vip male reproductive function, and Vax1^Vip males displayed an increase in total sperm through an unknown mechanism. These data support a theory that a less stringent circadian control may favor male fertility. Together our data, coupled with evidence from genetic knockout and light-disruption studies (92, 96, 97), indicate that male fertility is resilient when faced with circadian challenges.

In contrast to males, circadian disruption and impaired SCN function are known to have a strong impact on female reproduction (27, 98–100). VIP, in particular, has an important role in the female reproductive axis, where VIP knockout females have lengthened estrous cycles and are sub-fertile, resulting in fewer liters of smaller sizes (19). It is likely that some of the sub-fertility in full body VIP knockout females is driven by VIP neurons within the SCN, as our Vax1^Vip females also displayed lengthened estrous cycles. Our data are further supported by work in progress, showing that surgically ablated VIP neurons within the SCN also lengthened estrous cycles (101). Although Vax1^Vip estrous cycles were lengthened, we did not see changes in litter sizes or number of litters that are associated with full VIP knockout females. Taken together, these data indicate estrous cycle length is influenced by Vip-expressing neurons within the SCN.

The estrous cycle is regulated by hormonal feedback throughout the reproductive axis. Within the hypothalamus, VIP neurons directly project onto GnRH neurons to regulate the frequency of GnRH release to the pituitary (31) and indirectly through projections onto AVP neurons in the anteroventral paraventricular nucleus to regulate kisspeptin neurons that modulate the surge release of GnRH needed for ovulation (102, 103). Although only a single dose of kisspeptin was tested, we found that Vax1^Vip females had a greater release of LH in response to a kisspeptin challenge than Ctrls, a difference we did not observe in the males. The normal pituitary response to GnRH and normal circadian rhythms in ex vivo pituitary explants of Vax1^Vip females, combined with the absence of Vax1 in gonadotropes as shown by qPCR in isolated gonadotrope cells from female mice (53, 104) and single cell RNAseq [personal communication, (104, 105)], suggest it is unlikely that the increased LH release in response to kisspeptin is associated with abnormal gonadotrope function. One possibility for the sex difference in kisspeptin-induced changes in LH release in Vax1^Vip mice might be linked to differences in the neuronal circuit encompassing the sexually dimorphic anteroventral periventricular nucleus kisspeptin neurons (106). This neuronal population is larger in females than in males and plays a central role in the LH surge (107, 108). Interestingly, a comparable increased sensitivity to a kisspeptin challenge in females was also observed in full body Bmal1 knockout mice (8), where the mechanism for this increase remains unknown. Future work to determine how neuronal network changes, with or without intact molecular clock function, impact GnRH neuron sensitivity to kisspeptin will be of interest. Although the kisspeptin challenge elicited a greater LH release in the Vax1^Vip females than Ctrls, basal LH levels were comparable to Ctrls, and Vax1^Vip females exhibited a decrease in circulating FSH. LH and FSH production and release are controlled in a great part by the pulsatile pattern of GnRH release (87, 109, 110). Repeated blood sampling of Vax1^Vip mice would have been an ideal approach to assess for changes in pulsatile hormone release, however, as Vax1^Vip mice present with normal litter sizes, time to first litter, ovarian phase, which indicate overall normal ovarian function, and an increased activation of the stress axis, which is a suppressor of GnRH release (111), we decided against completing a LH and FSH pulse analysis due to the confounding effect of corticosterone on these data. In the future, we hope to better elucidate the relationship between stress and pulsatile hormones in this mouse model.

Changes in the neuroendocrine regulation of FSH release from the pituitary in Vax1^Vip mice is a potential pathway causing the reduction in estrogen of these mice. FSH is a limiting factor in the conversion of testosterone into estrogen in the granulosa cells of the ovary (112, 113), from where estrogen enters the general circulation. It is important to note that both FSH and estrogen are circadian (114, 115), thus a limit of this study, with blood sampling at a single time point, is our inability to assess if the reduction in these hormones might be due to a phase shift in hormone release. While primarily associated with its role in reproduction, estrogen has a multitude of functions, including altering the sensitivity of neuronal circuits (116, 117), regulating the activity of the stress axis (118), and displaying strong correlations with mood (119–121). Low estrogen has been correlated with depressive-like behaviors in women (122). In rodents, increasing estrogen has been shown to exert an anti-depressant-like effect during the Porsolt forced swim test, a test that is thought to reflect depressive-like behavior in rodents (123, 124). Additionally, an imbalance of progesterone and estrogen is associated with a higher incidence of mood disorders, including premenstrual dysphoric disorder (PMDD), a depressive disorder that presents with severe physical and physiological symptoms during the luteal phase of the menstrual cycle (125–128). In addition to recapitulating the low estradiol (or imbalance of progesterone and estrogen) as a risk factor for mood disorders, Vax1^Vip females also display another hallmark of PMDD, weakened SCN output (129, 130). Excitingly, Vax1^Vip females (but not males) have increased depressive-like behavior. Taken together, these data point to a novel role of VAX1 in regulating VIP neuron modulation of mood and the reproductive axis and raise the potential of Vax1^Vip females to serve as a new model for mood disorders that are tied to reproductive cycles, such as PMDD. Furthermore, VIP-and AVP-expressing neurons contribute to the regulation of the stress axis (131–134). Stressful situations result in an increase in signal from the hypothalamus, which translates to higher levels of corticosterone via activation of the hypothalamic-pituitary-adrenal (stress) axis. Notably, the hypothalamus is comprised of several nuclei, including the SCN (135) and the PVN (136, 137), with varying roles and contributions to the stress axis. Within the SCN, reductions in Vip have been correlated with increased stress (138), whereas the VIP neuron target, the PVN, is a central relay station of the stress axis (63, 139). Thus, the reduction of Vip in the SCN of Vax1^Vip females is a likely contributing factor in the increased corticosterone levels found in Vax1^Vip females, both at baseline and in response to a stressor. Specifically within the PVN, AVP is known to stimulate the stress axis (140, 141), and AVP expression in the PVN is comparable between control and VIP knockout males (142). This suggests that the reduction in VIP in the SCN of Vax1^Vip mice may not be driving the changes in Avp mRNA within the PVN, but more likely is the result of other VAX1 targets that impact VIP neuron communication with PVN neurons, such as GABA (143, 144).