The Transcription Factors Six3 and Six6 in Neuromedin‐S Neurons Differentially Affect Circadian Rhythms

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. Mice were maintained on a light/dark cycle of 12 h light, 12 h dark with light parameters as described previously (Van Loh et al. 2023). Six6^flox/flox mice (RRID:MMRRC_068240‐UCD; Pandolfi, Tonsfeldt, et al. 2019) were crossed with heterozygous Nms^cre mice (RRID:IMSR_JAX:027205; Lee et al. 2015) to generate mice discussed here as Six6^NMS. Similarly, Six3^flox/flox (RRID MGI:6507739; Lagutin et al. 2003) mice were crossed with Nms^cre mice to generate Six3^NMS mice; or Vip^cre mice (RRID:IMSR_JAX:010908, sourced from Jackson Labs) to generate Six3^VIP. In all conditional knockouts, Cre was maintained as heterozygous to avoid off‐target effects of the Cre allele (Joye et al. 2020). Cre negative, flox/flox littermates were used for controls (denoted Six3^fl/fl and Six6^fl/fl). The Ai14Rosa tdTomato reporter mice (RRID:IMSR_JAX:007914, sourced from Jackson Labs) were crossed with Nms^cre mice to create tdTomato^+/wtNms^cre (tdTomato^NMS mice) to visualize Cre‐containing neurons. Period2::Luciferase (PER2::LUC) mice were purchased from JAX (RRID:IMSR_JAX:006852) and crossed with Six6^NMS mice for at least two generations. The mice used for the resulting studies were Nms^creSix6^fl/fl PER2::LUC^+/−; control mice for these studies were Six6^fl/fl PER2::LUC^+/−. Genotyping primer sequences are in Table 1. Mice were screened for germline recombination using the primer combinations listed, and germline recombined animals were excluded from the studies.

Single‐cell RNA sequencing data were gathered using HypoMap (Steuernagel et al. 2022), a publicly available dataset using 17 published 10X Genomics and Drop‐seq mouse hypothalamus datasets. We utilized the CELLxGENE interface to probe the data. We sub‐selected SCN neurons following (Correa‐da‐Silva et al. 2025), first by selecting for neuronal cells, and then selecting for the SCN region. We then queried the set of 21,985 SCN neurons for expression of Nms, Six3, and Six6. Cells were considered positive if they had more than zero counts of the transcript.

Nms^cre male mice were mated with Rosa‐tdTomato^+/+ females and checked daily for vaginal plugs. Pups were retrieved on E16.5 and the heads drop‐fixed in 4% PFA overnight at 4°C. Tails were genotyped for Cre and Sry. After fixation, heads were sunk in 30% sucrose, embedded in O.C.T. (Tissue‐Tek, Sakura) and stored at −80°C until sectioning. Heads (n = 3, 2 female, 1 male) were sectioned coronally at 20 μm in two series. One series underwent gentle washing in PBS for 20 min followed by incubation in 1 μg/mL DAPI for 6 min, followed by another 20‐min PBS wash prior to coverslipping. Slides were imaged using a Zeiss Axio Imager M2. To show expression in adulthood, the brain from one 8‐week old mouse was drop‐fixed in 4% PFA as above. The brain was sectioned at 40 μm, and sections containing the SCN were mounted on slides. The SCN was visualized on an Andor BC43 (Oxford Instruments).

A detailed protocol on cell culture and transient transfections has been previously described (Meadows et al. 2022). In brief, mouse fibroblast NIH3T3 (RRID:CVCL_0594) cells were incubated in 10 cm plates in an incubator at 37°C with 5% CO₂ and maintained in complete media containing Dulbecco's Modified Eagles Media (DMEM), 10% fetal bovine serum (FBS), and 1% penicillin–streptomycin. Cells were transfected with a Six6 expression vector (Larder et al. 2011) or empty vector (pcDNA) and a reporter vector: Mouse Bmal1‐luc (a gift from Steven Brown [Addgene plasmid # 46824; RRID:Addgene_46824; Brown et al. 2005]) or Pgl empty vector; mouse Per2‐luc (Period 2; a gift from Joseph Takahashi [Addgene plasmid # 48746; http://n2t.net/addgene:48746↗; RRID:Addgene_48746; Yoo et al. 2005; Albrecht et al. 1997]) or Pgl2 empty vector; rat Avp‐luc (arginine vasopressin; a gift from Robert Shapiro 2000) or Pgl2 empty vector; or mouse Vip‐luc (vasoactive intestinal peptide; Hatori et al. 2014) or Pgl3 empty vector. To control for transfection efficiency, a reporter plasmid containing β‐galactosidase constitutively driven by the Herpes virus thymidine kinase promoter (TK‐βgal) was used.

On Day 1, cells were plated at 30,000 cells/well in a 12‐well plate and incubated overnight to allow cell adherence. On Day 2, cells were transfected with the following amounts of plasmid: TK‐βgal at 20 ng/well, luciferase plasmids or empty luciferase backbone at 400 ng/well, and pcDNA‐empty vector (EV) or Six6 expression vectors at 400 ng/well. Transfections were performed using the reagent Polyjet per manufacturer's instructions (SignaGen Laboratories). On Day 3, 24‐h post‐transfection the media was changed to fresh, complete media. On Day 4, 24 h post media change, the cells were harvested in lysis buffer (100 mM potassium phosphate pH 7.8 and 0.2% Triton X‐100) and plated in 96‐well plates and read for luciferase and β‐gal assays in a Luminometer microplate reader (Glomax, Promega). Within each well, luciferase values were normalized to TK‐βgal values, then triplicate luciferase/TK‐βgal values were averaged. Luciferase was normalized to the luciferase backbone for both Six6 and pcDNA. Fold change is displayed as normalized Six6/pcDNA.

Beginning at postnatal Day 21, pubertal onset was established by visual inspection of preputial separation (PPS) in males and vaginal opening (VO) in females, as described previously (Hoffmann 2018). Examination occurred between zeitgeber time (ZT) 2–4, and the assessor was blind to genotype. Body weight was recorded on the day pubertal onset was observed.

Estrous cyclicity was monitored by vaginal lavage with 20 μL H₂O or saline daily between ZT 3–5 for 15–18 consecutive days in 12–14‐week‐old mice. The lavage solution was dried on a glass slide and stained with 0.1%–0.5% methylene blue. Cytology was examined and scored by two independent, blinded observers. Smears were scored for the presence of leukocytes (diestrus), nucleated epithelial (proestrus), and cornified epithelial cells (estrus) (Byers et al. 2012). To assess fecundity, virgin 15–17‐week‐old female Six3^NMS or Six6^NMS mice were housed with 10–16‐week‐old male mice for 76 days. Mating cages were checked daily in the morning. Days to first litter, how many litters occurred in 76 days, and the number of pups per litter at birth were recorded.

Female mice underwent bilateral ovariectomy. Bosch et al. (2013), 4 days later mice received a subcutaneous dose of 0.25 μg estradiol (E2; Cat No. E8875, Sigma Aldrich) in sesame oil (Cat. No. S3547, Sigma Aldrich) between ZT 3–4. The following day, mice received a subcutaneous dose of 1.5 μg estradiol in sesame oil between ZT 3–4. The following evening (~30 h after injection), mice were sacrificed at lights out at ZT 12. Whole blood was collected, allowed to clot for 90 min, and spun for 20 min at 20,000 × g to separate serum. Serum was stored at −80°C until analysis. All procedures were performed blind to genotype.

LH was measured using the ultra‐sensitive LH assay (Kreisman et al. 2022). Briefly, serum samples were diluted 1:17 in assay buffer (PBS + 0.2% BSA). The sandwich ELISA was performed using LH 518B7 (UC Davis, Cat. no. 518B7, lot 13; RRID:AB_2665514) as a capture antibody at 1.0 μg/mL and biotinylated 5303 SPRN‐5 (Medix Biochemica RRID:AB_2784503) as a detection antibody at 4.379 μg/mL. A standard curve from 1 to 0.0019531 ng/mL was generated using Golden West Biosolutions, catalog No. TLIA1053.03. The plate was analyzed using a Tecan infinite 200 Pro plate reader at 490 nm minus the 650 nm background. A four‐parameter logistic curve was generated using MyAssays.com↗ website (http://www.myassays.com/↗). The intra‐assay CV was 7.22%, and the inter‐assay CV was 6.21%.

Testes, seminal vesicles, and epididymis were collected and weighed. Sperm was collected from the epididymis of male mice in 37°C M2 media (Sigma). 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 number of motile sperm was counted from a 10 μL aliquot of the M2 media containing sperm by using a hemocytometer. The hemocytometer was then placed on a heat block at 55°C for 5 min to immobilize all sperm. The total number of intact and immobile sperm was counted to determine percent motility. The second epididymis was cut into small pieces and left for 15 min at room temperature in M2 media. The solution was gently homogenized often to help liberate the sperm. The solution was filtered using a cell streamer (70 μm, Falcon), and sperm were diluted 1:10 with MQ before counting the total number of sperm heads.

To assess male plugging behavior, Six3^fl/fl or Six3^NMS male mice were paired with a virgin WT female for 10 days. During the assay period, the females and bedding were checked daily between ZT 2 and 3.

Pair‐housed mice were allowed to acclimate in chambers on a 12:12 light:dark cycle for 2 weeks. Tail bleeds were collected every 4 h for 24 h into capillary tubes, and serum was collected and stored at −20 C until assay. Approximately 10 μL of tail blood was collected using a capillary (Drummond Scientific) from each mouse at each timepoint. To minimize the number of times that the tail was cut for repeated blood collection, cuts from previous timepoints were re‐opened with a cotton swab soaked in saline. After the 24‐h collection in LD, mice were transitioned to constant darkness (DD) for 4 weeks before collection in DD. Because the activity rhythms of the mice were not monitored, tail bleeds in DD were collected based on clock time rather than the time relative to the activity onset of each animal. Corticosterone was measured from serum using the DetectX Corticosterone ELISA kit (Arbor Assays) according to the manufacturer's instructions. For all samples, 4 μL of serum was diluted 1:400 in dissociation reagent prior to loading on the assay. In some cases, time points from some individual animals were lost due to unanticipated collection issues. The plates were visualized using an iMark microplate reader (Bio‐Rad) and final concentrations were quantified using a four‐parameter logistic curve on MyAssays.com↗.

Female and male mice aged 8–12 weeks were singly housed in cages containing metal running wheels and wheel revolutions were monitored using magnetic sensors as previously described (Van Loh et al. 2023). Genotypes were randomly distributed in cages in a light‐tight cabinet with programmable lighting conditions. Rooms were monitored for temperature and humidity. 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) that integrated in 6 min 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), whereafter the mice were monitored for 4 weeks in constant darkness (DD), with wheel running data analyzed from Weeks 2–4 (14 days) in DD. In some cases, mice were exposed to a 2 week “skeleton” light paradigm (1 h light: 11 h dark: 1 h light: 11 h dark) in between LD and DD. We found no evidence of masking in the skeleton photoperiod phase angle between Six3^fl/fl and Six3^NMS (−0.05 ± 0.50 vs. −0.04 ± 0.38 h, t(11) = −0.036, p = 0.972, n = 6–7) or Six6^fl/fl and Six6^NMS (0.37 ± 0.13 vs. 0.23 ± 0.20 h, t(16) = 1.688, p = 0.114, n = 8) so we excluded skeleton photoperiods from future experiments. Cage changes were scheduled for 2–4‐week intervals. Light intensity varied between 268 and 369 l× inside the mouse cage with the wheel. Wheel running activity was analyzed using ClockLab Analysis (ActiMetrics) by an experimenter blind to experimental group. Circadian period (tau, or free‐running rhythm) was analyzed by constructing a least‐squares regression line through a minimum of 14 daily activity onsets in constant darkness. Rhythm strength was assessed using the amplitude of a χ² periodogram from the first 2 weeks of LD or the final 2‐week period of DD. Alpha was determined based on the length of time that activity was greater than 20% of the peak activity. Onset precision was calculated as the standard deviation of onsets compared to the time of lights on (LD) or the predicted tau (DD). Mice were compared to their littermate controls, and male and female mice were grouped for analysis except where noted.

For circadian rhythm organotypic explant studies, tissues from mice expressing the PER2::LUC circadian reporter were collected and analyzed as previously described (Yaw et al. 2020). Proestrus Six6^NMS PER2::LUC and PER2::LUC control females were placed on LD and euthanized at ZT 3–4. The brain, pituitary, ovaries, and uterus were removed immediately and placed in an ice‐cold, CO₂ saturated Hank's Balanced Salt Solution (HBSS) for approximately 1 h. Using a Vibratome (Leica), tissue sections of 300 μm were collected and the indicated brain region was dissected from the slices in ~2 × 2 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 ZT 6–6.5, and recordings were started. The bioluminescence was counted for 70 s every 10 min 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 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). Period was defined as the time in hours between the peaks of the fitted curve.

Animals were sacrificed between ZT 8–11. Brains were flash frozen, sectioned at 20 μm, and stored at −80°C. Slides containing the SCN underwent multiplex in situ hybridization detection of mouse ( Mus musculus ) mRNAs with RNAscope LS Multiplex Fluorescent Reagent Kit (Advanced Cell Diagnostics) for 3‐plex assay in addition to RNAscope LS 4‐Plex Ancillary Kit (Advanced Cell Diagnostics) for 4‐plex assay following the vendor's standard protocol for FFPE tissue sections with minor modifications. RNAscope assays were performed on a Leica Bond autostainer as previously described (Sempere et al. 2020; Van Loh et al. 2023) with the following probes: 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 and tissue slides were counterstained with DAPI. Sections were imaged using a Zeiss Apotome.2 equipped with a Zeiss AxioCam 506 (Zeiss, Germany). Qupath (Bankhead et al. 2017) was used to automate cell detection based on DAPI staining. Cells were classified based on detection of subcellular objects indicating Avp, Vip, and Nms mRNA. Results were obtained per single SCN hemisphere and averaged by animal (2–4 hemispheres per animal).

Statistical analyses were performed with GraphPad Prism 10. Sample sizes were determined based on previous studies in our laboratories or pilot studies. Outliers were identified using ROUT analysis (Q = 1%). Data were analyzed using unpaired t‐test, Mann–Whitney U‐test, one‐way ANOVA or two‐way ANOVA, followed by post hoc analysis where appropriate. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

SIX3 and SIX6 are highly expressed in the adult SCN (Hatori et al. 2014; VanDunk et al. 2011), but the overlap with each other and NMS is unknown. Using published single‐cell RNA sequencing data from HypoMap containing 17 individual scRNA datasets of mouse male and female hypothalamus (Steuernagel et al. 2022), we examined the overlap between Nms, Six3, and Six6 expression in 21,985 SCN neurons. We found that 28% of SCN neurons express Nms, consistent with previous reports (Lee et al. 2015). Of Nms + neurons, 78% express Six3, 62% express Six6, and 52% express both Six3 and Six6 (Figure 1a). The cells that contain Nms, Six3, and Six6 represent approximately 14% of SCN neurons. To compare our results with the previous findings in the Six3^fl/fl‐Synapsin^cre mouse, we also evaluated Syn1 expression in these neurons and found that a larger population of SCN neurons express Syn1 (37%), that 56% of Nms + neurons express Syn+, and 67% of Syn1+ neurons express Nms. These results indicate that the Nms^cre targets a distinct population of the SCN compared to Synapsin^cre. Having established that approximately half of Nms neurons express Six3 and Six6, we then wanted to determine the developmental timing of Nms^cre expression in the SCN. We used the Nms^cre mice crossed with a Cre‐dependent tdTomato reporter to visualize neurons expressing Cre, as the tdTomato signal is expressed following Nms^cre‐mediated recombinase activity. To determine if Nms^cre turns on after SCN neurogenesis, we examined the SCN of E16.5 tdTomato^NMS embryos (Figure 1b,c). We observed no tdTomato‐containing cells in the SCN in either males or females (n = 3, 2 females), in contrast to robust tdTomato expression in the adult SCN (Figure 1d). We did observe tdTomato‐containing populations lateral to the SCN, indicating that there are areas of Nms^cre expression at E16.5. We concluded that the activation of Nms^cre, and presumably NMS expression, occurs subsequent to SCN neurogenesis (~E12‐E15) (Kabrita and Davis 2008; Shimada and Nakamura 1973; VanDunk et al. 2011). Therefore, targeting Six3 and Six6 using Nms^cre would allow us to isolate only those effects which occur after SCN neuron cell proliferation is complete and which are unlikely to influence the development of the SCN itself.

SIX3 can regulate molecular clock and SCN peptides in heterologous systems (Hoffmann et al. 2021; Meadows et al. 2022). SIX3 and SIX6 are highly homologous, and we hypothesized that SIX6 would have similar promoter activity on clock genes and neuropeptides as SIX3. Six6 overexpression significantly increased expression of the Bmal1 luciferase reporter (Mann–Whitney Test, U = 0, p = 0.002; Figure 2a), which expresses the region ~1 kb upstream of Bmal1, an in‐frame coding region and ~1 kb of the Bmal1 3′ UTR (Brown et al. 2005). The Per2 promoter expresses −1128 to +2129 of the promoter/enhancer region (Yoo et al. 2005), and was significantly repressed after 24 h (Mann–Whitney test, U = 6, p = 0.005; Figure 2b). We also found that Six6 significantly increased Vip reporter activity (Mann–Whitney test, U = 12, p = 0.038; Figure 2c), which reflects a 1 kb promoter region upstream of Vip (Hatori et al. 2014). Six6 did not significantly regulate a 5.5 kb region of the rat Avp promoter (Mann–Whitney test, U = 15, p = 0.083; Figure 2d) (Shapiro 2000). Together, these data demonstrate that, like SIX3, SIX6 can regulate the promoters of genes critical for SCN function.

Luteinizing hormone (LH) is vital for proper pubertal development, ovulation, and maintenance of fertility. NMS has previously been shown to interact with LH in rats (Vigo et al. 2007), and Six3^fl/fl‐Synapsin^cre females have reduced LH and delayed puberty (Hoffmann et al. 2021). To determine how loss of Six3 and Six6 in NMS neurons affects fertility, we assessed female age at vaginal opening (Table 2), estrous cycle length (Figure 3a–d), time to first litter (Figure 3e,f), number of litters in 76 days (Figure 3g,h), and pups per litter (Figure 3i,j). We found no differences between Six3^NMS females and Six3^fl/fl controls in estrous cycle length (unpaired t‐test, t(21) = 0.474, p = 0.640), time to first litter (Welch's t‐test, t(11.72) = 0.680, p = 0.510), number of litters (unpaired t‐test, t(17) = 0.275, p = 0.787), or pups per litter (unpaired t‐test, t(18) = 0.477, p = 0.642). Similarly, we found no difference between Six6^NMS females and Six6^fl/fl controls in estrous cycle length (Welch's t‐test, t(8.338) = 1.867, p = 0.097), time to first litter (Welch's t‐test, t(5.729) = 1.275, p = 0.252), number of litters (unpaired t‐test, t(14) = 1.755, p = 0.101), or pups per litter (unpaired t‐test, t(14) = 0.439, p = 0.667).

We then explored whether loss of Six3 or Six6 in NMS neurons could disrupt the timing of the preovulatory LH surge, which is regulated by the SCN. Using an induced surge model in ovariectomized mice, mice were sacrificed during the morning (ZT 2) or at lights off (ZT 12, expected time of LH surge). The surge threshold was the average of the AM control LH values plus three standard deviations (Figure 3k). Six3^fl/fl and Six6^fl/fl mice were combined for the control group. We found that 6 of 7 control mice exhibited a PM surge, along with all of the Six6^NMS (4/4) and Six3^NMS (5/5) mice. LH levels indicated a significant effect of time (as expected), but not of genotype (two‐way ANOVA, time F(1, 24) = 30.36, p < 0.001, genotype F(2, 24) = 2.223, p = 0.130, interaction F(2, 24) = 1.695, p = 0.205). Therefore, the ability to mount a normally timed LH surge is unaffected by the loss of either Six3 or Six6 in NMS neurons in LD conditions.

To determine whether loss of Six3 or Six6 affects the hypothalamic–pituitary‐gonadal axis in males, we first observed pubertal onset in both Six3^NMS and Six6^NMS male mice. No differences were found in the weight or age at preputial separation between either Six3^NMS or Six6^NMS mice and respective flox/flox controls (Table 2). After sexual maturity, we measured total sperm (Figure 4a,b), and the percent motile sperm (Figure 4c,d) in both Six3^NMS and Six6^NMS male mice. Six3^NMS males showed a reduction in the percent of motile sperm (unpaired t‐test, t(10) = 2.632, p = 0.025). We measured plugging efficiency in Six3^NMS mice by pairing them with wild‐type females and measuring the time to plug and found no difference compared to control Six^fl/fl males (2.38 ± 0.59 vs. 2.87 ± 0.44 days, unpaired t‐test, t(14) = 0.675, p = 0.511). Then, we paired male Six3^NMS or Six6^NMS with wild‐type females and examined the number of pups per litter (Figure 4e,f) and found no differences in number from either line (unpaired t‐test, t(10) = 1.001, p = 0.340 and t(15) = 0.128, p = 0.900, respectively).

A possible mechanism for reduced sperm motility in Six3^NMS is a local testis effect. NMS has previously been reported in the testis of rats (Ida et al. 2005), and Six3 is expressed in spermatids and spermatocytes (Lukassen et al. 2018b, 2018c). We examined whether Nms^cre was expressed in the test is using a tdTomato^NMS reporter mouse. In mature males, we found evidence of sporadic tdTomato expression, raising the possibility that a local testis knockdown of Six3 may drive the phenotype (Figure 4g).

We monitored corticosterone rhythm in Six3^NMS mice to determine if loss of Six3 in NMS‐containing neurons impacted the circadian rhythm of corticosterone. After habituation, male and female mice in 12 h light:12 h dark (LD) conditions were tail bled every 4 h for 24 h. Using a two‐way ANOVA analysis, corticosterone levels showed a significant effect of time in both groups, and no effect of genotype (two‐way repeated measures ANOVA, effect of time F(5, 155) = 10.110, p < 0.001; genotype F(1, 31) = 0.000066, p = 0.994; interaction F(5, 155) = 0.855, p = 0.513; Figure 5a). Mice were then allowed to free run for 2 weeks in constant dark (DD) conditions before undergoing tail bleeds again. Sampling times were not aligned to circadian activity, and due to drifts in the corticosterone peak due to individual differences in free running rhythms, the data are shown as aligned to the peak cortisol time (Figure 5b). Similar to the LD conditions, we found a significant effect of time and no effect of genotype (mixed‐effects analysis, time F(3.33, 95.91) = 28.74, p < 0.001; genotype F(1, 29) = 0.1668, p = 0.686; interaction F(3.33, 95.91) = 2.049, p = 0.106). Therefore, Six3 in NMS neurons does not contribute to the amplitude or timing of the corticosterone surge.

To determine how SCN function was altered in Six3^NMS and Six6^NMS mice, we measured wheel running locomotor activity. First, we wanted to confirm that the presence of the Nms^cre allele alone had no effect on locomotor activity. We found that the free‐running rhythm was unaffected by the Nms^cre allele (Table 3; Figure 6a,f; unpaired t‐test, t(27) = 0.426, p = 0.674). We also evaluated circadian amplitude, onsets precision, phase angle, and alpha, and found no effect by unpaired t‐test (Table 4). After establishing no effect of the Nms^cre allele itself, we then asked whether conditional knockout of Six6 in NMS neurons (Six6^NMS) would affect locomotor activity compared to littermate Six6^fl/fl controls. We found no significant difference in the free running period (Figure 6b,g; unpaired t‐test, t(25) = 0.166, p = 0.869) or other locomotor behavior (Table 3). We confirmed this lack of circadian change with ex vivo explants from the SCN, arcuate nucleus, pituitary gland, ovaries, and uterus, which showed no significant differences in the period of the PER2::LUC rhythms between PER2::LUC‐Six6^NMS mice and controls (Figure 6c; Table 4).

Finally, we wanted to examine the effect of conditional loss of Six3 in SCN populations. We measured locomotor activity in both Six3^NMS mice as well as Six3^fl/fl x Vip^cre mice (Six3^VIP) neurons, given the extensive overlap of VIP and NMS (Lee et al. 2015). We found that there was a significant difference in the free running rhythm between Six3^fl/fl, Six3^NMS, and Six3^VIP (Figure 6d,g; one‐way ANOVA (F(2, 62) = 16.34, p < 0.001)). Post hoc analysis indicated that both Six3^NMS and Six3^VIP were significantly different from Six3^fl/fl controls (Šidák's multiple comparisons test, p < 0.001 and p = 0.004, respectively). There were no other differences by one‐way ANOVA among the other measures (Table 3).

To probe a potential mechanism for the accelerated free running rhythm of Six3^NMS mice, we hypothesized that these animals may have an alteration in SCN peptides. We performed RNAscope for Avp, Vip, and Nms in the SCN of mice harvested from ZT 8–11 (Figure 7a, n = 3–5). We found that there was no significant difference in the percentage of neurons that express Vip between Six3^fl/fl and Six3^NMS mice (unpaired t‐test, t(6) = 0.620, p = 0.558; Figure 7b), Avp (unpaired t‐test, t(6) = 0.948, p = 0.380; Figure 7c), or Nms (unpaired t‐test, t(6) = 0.921, p = 0.392; Figure 7d).

The data that support the findings of this study are available from the corresponding author upon reasonable request.