Circadian regulation and molecular role of the Bsx homeobox gene in the adult pineal gland

Abstract

The pineal gland is a neuroendocrine organ responsible for production of the nocturnal hormone melatonin. A specific set of homeobox gene-encoded transcription factors govern pineal development, and some are expressed in adulthood. The brain-specific homeobox gene (Bsx) falls into both categories. We here examined regulation and function of Bsx in the mature pineal gland of the rat. We report that Bsx is expressed from prenatal stages into adulthood, where Bsx transcripts are localized in the melatonin-synthesizing pinealocytes, as revealed by RNAscope in situ hybridization. Bsx transcripts were also detected in the adult human pineal gland. In the rat pineal gland, Bsx was found to exhibit a 10-fold circadian rhythm with a peak at night. By combining in vivo adrenergic stimulation and surgical denervation of the gland in the rat with in vitro stimulation and transcriptional inhibition in cultured pinealocytes, we show that rhythmic expression of Bsx is controlled at the transcriptional level by the sympathetic neural input to the gland acting via adrenergic stimulation with cyclic AMP as a second messenger. siRNA-mediated knockdown (>80% reduction) in pinealocyte cultures revealed Bsx to be a negative regulator of other pineal homeobox genes, including paired box 4 (Pax4), but no effect on genes encoding melatonin-synthesizing enzymes was detected. RNA sequencing analysis performed on siRNA-treated pinealocytes further revealed that downstream target genes of Bsx are mainly involved in developmental processes. Thus, rhythmic Bsx expression seems to govern other developmental regulators in the mature pineal gland.

Introduction

The mammalian pineal gland secretes the nocturnal hormone melatonin, which is synthesized from tryptophan by the action of the pineal enzymes tryptophan hydroxylase 1 (TPH1), arylalkylamine N-acetyltransferase (AANAT), and acetylserotonin O-methyltransferase (ASMT)¹. Rhythmic melatonin synthesis in the pineal gland is controlled by the circadian clock of the hypothalamus, i.e. the suprachiasmatic nucleus, via sympathetic nerve fibers originating from the superior cervical ganglia (SCG) which release norepinephrine (NE) in the gland at night². This rhythmic adrenergic input induces transcription and activation of AANAT resulting in nocturnal melatonin synthesis^1,3.

Development of the pineal gland is controlled by a specific set of homeobox gene-encoded transcription factors^4–6. However, as a special feature of the pineal gland, a number of homeobox genes continue to be expressed into adulthood or are even restricted to the postnatal gland^4,5,7–9. To this end, we have shown that certain pineal homeobox genes are involved in transcriptional regulation of the enzymes of the melatonin synthesis pathway and that rhythmic expression of these transcription factors is also controlled by the adrenergic input to the gland^10–12.

The brain-specific homeobox gene (Bsx) is expressed in the developing pineal gland at early stages and is essential for proper pineal morphogenesis in both mice, frogs (Xenopus) and zebrafish^13–16. In the postnatal rodent brain, Bsx is specifically expressed in the pineal gland and the hypothalamus¹⁴. Large screening efforts have previously identified Bsx as being expressed in the mature rat pineal gland with high levels during nighttime and regulated by sympathetic adrenergic signaling¹⁷; in addition, RNA sequencing studies have detected adult pineal expression of Bsx in zebrafish, chicken, and several mammals (https://snengs.nichd.nih.gov). However, circadian and developmental expression patterns, cellular localization, and regulatory functions of Bsx in the postnatal mammalian pineal gland remain to be established. In this study, we employed histological and molecular biological techniques to identify the role of Bsx in the adult pineal gland.

Materials and methods

Animals

Sprague-Dawley rats (timed-pregnant mothers or 180 to 220g adult males) obtained from Charles River (Sulzfeld, Germany) or from Janvier Labs (Le Genest-Saint-Isle, France) were housed under a 12h light: 12h dark (12L:12D) schedule. Euthanasia was performed at indicated Zeitgeber times (ZT) or circadian times (CT) by anesthesia with CO₂ followed by decapitation; during dark conditions, euthanasia was carried out under dim red light. For the developmental series, heads (embryonic day 15 (E15) to postnatal day 2 (P2)) or whole brains (P10 to P30) were removed, immersion fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) for 2 days at 4°C, cryoprotected in 25% sucrose, and frozen on crushed solid CO₂. For qRT-PCR and in situ hybridization on adult tissue, whole brains, brain parts and peripheral tissues were dissected and immediately frozen on crushed solid CO₂. For circadian experiments, animals were transferred to constant darkness (DD) two days prior to euthanasia. Superior cervical ganglionectomy (SCGx) and decentralization of the superior cervical ganglia (SCGdcn) were performed as previously described¹⁸. Injection (i.p.) of isoprenaline (10 mg/kg in phosphate buffered saline (PBS)) or PBS was done at ZT5 followed by euthanasia at ZT8. For pineal gland cell cultures, pineal glands dissected from adult male rats (n=12–22, ZT3–5) were pooled and kept in cold high-glucose Dulbecco’s Modified Eagle Medium (DMEM) until pineal cell isolation. All animal experiments were in accordance with the guidelines of EU directive 2010/63/EU and approved by the Danish Council for Animal Experiments (authorization number 2017–15-0201–01190) and the Faculty of Health and Medical Sciences, University of Copenhagen (authorization number P17–311).

Human tissue

Human adult brain sections were obtained from the Human Brain Tissue Bank of Semmelweis University (Budapest, Hungary): #233, pineal gland, male, age 54, time of death 4.25 AM, heart failure; #249, pineal gland, male, age 22, time of death 23.10 PM, suicide; #273, hypothalamus, male, age 64, time of death unknown, stroke. Procedures were approved by the Regional and Institutional Committee of Science and Research Ethics of Scientific Council of Health (34/2002/TUKEB-13716/2013/EHR, Budapest, Hungary) and the Code of Ethics of the World Medical Association (Declaration of Helsinki).

Radiochemical in situ hybridization

Cryostat brain sections (adult rats, 12 µm; developmental series, 14 µm; human tissue, 18 µm) were hybridized with [³⁵S]dATP-labelled antisense DNA probes specific for human or rat Bsx mRNA (Table 2), as previously described¹⁹. Hybridized sections were washed, exposed to an X-ray film for 16 to 21 days, and developed. The X-ray films were digitized; optical densities were quantified using Scion Image Beta 4.0.2 (Scion, Frederick, MD, USA) and converted to dpm/mg tissue by using simultaneously exposed ¹⁴C-standards. Sections were counter-stained in cresyl violet.

Table 2.

Radiochemical and RNAscope in situ hybridization probes.

Transcript	Species	Genbank acc. #	Technique	Position	Probe type / sequence (5´−3´)
Bsx	Rat	NM_001191995.1	Radiochemical	76-42	GATGGAGAGGGGAAGTGAAGTTGAGATTCATCTTG
			RNAscope	2-856	18ZZ probe set
	Human	NM_001098169.1	Radiochemical	254-220	GTGAGGAAATAAGGATGATGGTGGTCTCCCTTATG
Aanat	Rat	NM_012818.2	RNAscope	23-1176	20ZZ probe set

RNAscope in situ hybridization

Fresh frozen cryostat brain sections (adult rats, 12 µm) were pretreated and hybridized in accord with the manufacturer’s instructions (ACDBio, Newark, CA; protocols #320513 and #320293). RNAscope in situ hybridization²⁰ was performed by use of the RNAscope Fluorescent Multiplex Reagent Kit (ACDBio) with probe sets specific for Bsx and Aanat mRNA (Table 2). Sections were photographed in a Zeiss AxioImager Z1 epifluorescence microscope equipped with a Zeiss AxioCam 506 mono camera.

Brightness and contrast were adjusted in Adobe Photoshop 7.0 (Adobe Systems Software, San Jose, CA).

Quantitative real-time Reverse Transcriptase PCR (qRT-PCR)

Total RNA was isolated using TRIzol (Life Technologies, Carlsbad, CA) followed by DNase I treatment (Invitrogen, Taastrup, Denmark). cDNA was synthesized following the Superscript III protocol (Invitrogen) using 500 ng (from tissues) or 250 ng (from cell cultures) of total RNA as starting material. PCR reactions were run in a Lightcycler 96 (Roche Diagnostics, Hvidovre, Denmark) at volumes of 10 µl containing Faststart Essential DNA Green Master (Roche), 0.2 µl cDNA, and 0.5µM transcript-specific primers (Table 1), using the following program: 95°C for 10 minutes; 40 cycles of 95°C for 10 seconds, 63°C for 10 seconds, 72°C for 15 seconds. Product specificity was initially confirmed by melting curve analysis and gel electrophoresis. 10-fold serial dilutions of pUC57 plasmids (Genscript, Piscataway, NJ) containing the target sequence were used to generate standard curves. Copy numbers were normalized against those of the housekeeping transcript Gapdh or geometric means of Gapdh and Actb copy numbers.

Table 1.

qRT-PCR primer sequences.

Transcript	Genbank acc. #	Position	Forward primer sequence (5´−3´)	Reverse primer sequence (5´−3´)
Aanat	NM_012818.2	525-682	TGCTGTGGCGATACCTTCACCA	CAGCTCAGTGAAGGTGAGAGAT
Actb	NM_031144.3	26-184	ACAACCTTCTTGCAGCTCCTC	CCACGATGGAGGGGAAGAC
Asmt	NM_144759.2	696-860	GCAAGACCCAGTGTGAGGTT	CAGTAGTGCACCACCTGGC
Bsx	NM_001191995.1	231-416	CACCCTCCTCACCCCTCACAC	CCGGAGAGCTGCGAGTCAGAA
Crx	NM_021855.1	76-204	ATGCACCAGGCTGTCCCATA	TGCATACACATCCGGGTACTG
Gapdh	NM_017008.4	77-386	TGGTGAAGGTCGGTGTGAACGGAT	TCCATGGTGGTGAAGACGCCAGTA
Otx2	NM_001100566.1	630-761	ACCCAGACATCTTCATGCGG	TCTGACCTCCGTTCTGTTGC
Pax4	NM_031799.1	674-808	TGCCTGAAGACACAGTGAGG	TGATCCCTGGAGAATCTTTTGGT
Rax	NM_053678.1	461-581	GTGCCTTTGAGAAGTCCCACT	CGTCTCCACTTGGCTCGAC
Tph1	NM_001100634.2	702-814	CAGAAACCTTCCTCTGCTCTCA	CAGGACGGATGGAAAACCCT

Pineal cell cultures and siRNA-mediated knockdown

Primary pineal cell cultures were prepared as previously described ¹⁰ and incubated in aliquots of 100,000 cells per sample. For knockdown experiments, pinealocytes were transfected with siRNA specifically targeting Bsx mRNA (siBsx) (40 nM; Thermo Fisher Scientific; Stealth RNAi Bsx siRNA, catalogue #1330001, targeting position 182–206 on NM_001191995.1) or non-targeting siRNA (siNT) (40 nM; Thermo Fisher Scientific; Stealth RNAi siRNA negative control, catalogue #12935300) using Lipofectamine RNAiMAX Reagent (Thermo Fisher Scientific). On the second day, the medium was changed. On the third day, norepinephrine (NE) (3 µM; Sigma-Aldrich, Søborg, Denmark) or dibutyryl-cyclic adenosine monophosphate (DBcAMP) (500 µM; Sigma-Aldrich) were added to the cell media for the indicated durations. For transcriptional inhibition, pineal cell cultures were pretreated with Actinomycin D (30 µg/ml; Sigma-Aldrich) for 1h prior to DBcAMP stimulation. Harvested cells were immediately frozen on crushed solid CO₂ and stored at −80°C.

Statistical analyses

Quantitative data analysis of qRT-PCR and in situ hybridization data was performed in Graphpad Prism 8.00 (GraphPad Software, La Jolla, CA). Data were analyzed by one- or two-way ANOVA followed by Bonferroni-corrected multiple individual t-tests for the relevant comparisons; individual t-tests were performed as unpaired Student’s or Welch’s t-tests depending on whether the variances of two compared groups could be presumed equal, as determined by F-tests.

RNA sequencing

Total RNA from pineal cell cultures was isolated using TRIzol (Life Technologies) followed by purification and DNase I treatment using a Qiagen RNeasy Micro Kit (Qiagen, Germantown, MD). RNA amount and integrity were assessed using a Bioanalyzer 2100 (RIN > 8) (Agilent, Santa Clara, CA); total input was 3.4 ng. cDNA libraries were prepared with Takara SMARTer Stranded Total RNAseq Pico Input Mammalian Kit v2 with unique barcode adapters for each of the six libraries according to the manufacturer’s instructions (Takara Bio, Mountain View, CA). Sequencing was performed on a single FlowCell on an Illumina HiSeq 2500 (Illumina, San Diego, CA), resulting in an average of 68 million 2x100 bp reads per sample. The number of reads per transcript was normalized to the total number of reads; reads were aligned to the rat genome (RefSeq rn6) using the RNA STAR algorithm²¹. Differential expression analysis was performed in DESeq2 software using Wald tests with Benjamini-Hochberg corrections for False Discovery Rate²². Confidently differentially expressed genes were included in an analysis for transcript overrepresentation using the DAVID gene-annotation enrichment tool where we examined the GOterm Biological Processes at level 5 output after filtering for the following parameters: P<0.001, ≥±25% relative change, base mean ≥100 reads^23,24.

Results

Bsx is expressed in the developing rat pineal gland into adulthood

To examine the ontogenetic expression patterns of Bsx, in situ hybridization was performed on coronal brain sections from rats sacrificed at daytime (ZT6) and at nighttime (ZT18) during pre- and postnatal stages ranging from E15 to P30 (Fig. 1). Bsx was detected in the pineal gland at all ages, as reported in other species^13,15,16, but was undetectable in other brain regions present in the same coronal plane. Both developmental age and ZT exerted significant effects on Bsx expression in the pineal gland (P<0.0001) (Fig. 2).

Figure 1. Ontogenetic expression of Bsx in the rat brain.

Representative X-ray images of coronal brain sections hybridized for detection of Bsx transcripts at daytime (ZT6, middle) or nighttime (ZT18, right) at the indicated developmental ages. Images of cresyl violet-counterstained brains at ZT6 (left) are shown for comparison. Scale bar, 1 mm. E, embryonic day; P, postnatal day; ZT, Zeitgeber time.

Figure 2. Quantification of Bsx expression in sections of the developing rat brain.

Densitometric analysis of radiochemical in situ hybridization for detection of Bsx transcripts in the developing pineal gland of rats at daytime (ZT6, dashed line) or nighttime (ZT18, solid line) at the indicated ages. Two-way ANOVA revealed significant effects of both developmental age (P<0.0001) and time of day (P<0.00001) on expression of Bsx in the pineal gland. Values on the graph represent mean ± SEM; n=4. ZT, Zeitgeber time.

Bsx is expressed in melatonin-producing pinealocytes of the adult rat pineal gland and exhibits a circadian rhythm

To determine the tissue distribution of Bsx transcripts, qRT-PCR was performed on samples from the central nervous system and peripheral tissues obtained from adult rats sacrificed at ZT6 and ZT18 (Fig. 3A). Bsx mRNA was detectable in the pineal gland and the hypothalamus^{13,14,25–27}; significantly lower levels were detected in the hypothalamus (P<0.001). Pineal Bsx mRNA showed a significant day-night difference with higher levels at ZT18 compared to ZT6 (P<0.05), as previously indicated by RNA sequencing studies¹⁷. A day-night difference in hypothalamic Bsx expression was not detected (P>0.05). In situ hybridization on adult rat brain sections confirmed tissue-specific expression in the pineal gland (Fig. 3B). BSX mRNA was further detected in sections of the human pineal gland with a positive signal in the parenchyma; a signal above background was not detected in sections of the hypothalamus containing the supraoptic nucleus, which was used as negative control (Fig. 3C)^13,14. Cellular distribution of Bsx in the rat pineal gland was examined by RNAscope in situ hybridization; cells expressing Bsx were present in all parts of the gland and the Bsx signal colocalized with the Aanat signal, showing that Bsx is expressed in melatonin-producing pinealocytes (Fig. 4)

Figure 3. Tissue-specific expression of Bsx in the adult rat and human pineal gland.

A) Quantitative analysis of Bsx transcript levels as determined by qRT-PCR in adult rat tissues obtained at daytime (ZT6) or nighttime (ZT18). Bsx copy numbers were normalized against geometric means of the copy numbers of the housekeeping genes Gapdh and Actb, which were detected in all examined tissues. Values on the Y-axis are shown on a logarithmic scale. Values on the graph represent mean ± SEM; n=3, except for the pineal gland, where n=6 (ZT6) and n=5 (ZT18). Tissue distribution (P<0.001) and ZT (P<0.0001) both exerted significant effects on Bsx mRNA levels in the tissues where it was detected as determined by two-way ANOVA. P-values determined by Bonferroni-corrected Welch’s t-tests: ns, not significant; *, P<0.05. B) Radiochemical in situ hybridization on coronal (left) and sagittal (right) sections for detection of Bsx transcripts in the rat brain. Animals were sacrificed at ZT18. Scale bars, 1 mm. C) Radiochemical in situ hybridization for detection of BSX expression in coronal sections of the human pineal gland (left, top) and the hypothalamus (left, bottom). Images of the same sections counterstained with cresyl violet are shown for comparison (right). Sections shown are from #249 (pineal gland) and #273 (hypothalamus). Scale bar, 1 mm. 3V, third ventricle; SON, supraoptic nucleus; ZT, Zeitgeber time.

Figure 4. Cell-specific Bsx expression in the adult rat pineal gland.

RNAscope in situ hybridization for detection of Bsx (red) and Aanat (orange) transcripts in coronal sections of the pineal gland of rats sacrificed at Zeitgeber time 18. Sections were also stained in DAPI for comparison. Arrow heads indicate the same cell photographed with different filter settings. Scale bars, 100 µm (left) and 20 µm (right).

Radiochemical in situ hybridization on sections from brains obtained throughout the day showed a significant 10-fold diurnal rhythm in Bsx mRNA levels of the pineal gland in 12L:12D (P<0.01) (Fig. 5A,C). A similar rhythmic expression was detected in DD (P<0.0001), showing that pineal Bsx expression is circadian in nature (Fig. 5B,C). Under both lighting conditions, pineal Bsx expression reached a peak in the middle of the night (Fig. 5C).

Figure 5. Rhythmic expression of Bsx in the adult rat pineal gland.

A, B) Radiochemical in situ hybridization for detection of Bsx transcripts in the pineal glands of rats sacrificed at the indicated time points under a 12h light:12h dark schedule (12L:12D) (A) and in constant darkness (DD) (B). Scale bars, 1 mm. C) Densitometric quantification of radiochemical in situ hybridization corresponding to A and B. One-way ANOVA detected significant effects of time on Bsx transcript levels in both 12L:12D (P<0.01) and DD (P<0.0001). Fold changes (ratios between highest and lowest measured transcript levels) were 10.0 ± 6.5 for Zeitgeber time (ZT) 15:ZT3 and 11.3 ± 1.2 for circadian time (CT) 15:CT6, while cosinor analysis estimated the times of peak expression at ZT18.8 ± 2.1 in 12L:12D and CT17.5 ± 2.0 in DD. Shaded areas indicate dark conditions; breaks in the light/dark bar indicate that sampling was performed separately for LD and DD. Values on the graph represent mean ± SEM; n=3, except for ZT3 (n=4), ZT12 (n=4), and ZT15 (n=2). CT, circadian time; ZT, Zeitgeber time.

Pineal Bsx expression is driven by sympathetic adrenergic signaling acting at the transcriptional level

The nocturnal expression of Bsx and previous large-scale RNA sequencing analyses indicate that pineal Bsx is regulated by the sympathetic adrenergic input to the gland via the second messenger cAMP¹⁷. The neural pathway connecting the suprachiasmatic nucleus with the pineal gland was lesioned by either decentralizing the SCG (SCGdcn), i.e. cutting the preganglionic fibers in the sympathetic trunk leading to the SCG, or removing the ganglia (SCGx). In situ hybridization on the pineal gland showed that these surgical procedures significantly affected pineal Bsx expression (P<0.0001). Bsx mRNA levels were significantly higher at ZT18 in control rats (P<0.05), while no differences in Bsx mRNA levels were detected between day and night in neither SCGx nor SCGdcn rats (P>0.05), showing that an intact neural pathway is necessary for maintaining daily changes in pineal Bsx expression (Fig. 6A).

Figure 6. Regulation of pineal Bsx expression by the sympathetic adrenergic input to the gland.

A) Radiochemical in situ hybridization for detection of Bsx transcripts in the pineal gland of rats with superior cervical ganglionectomy (SCGx), decentralized superior cervical ganglia (SCGdcn), and corresponding controls. Two-way ANOVA detected significant effects of both ZT (P<0.001) and surgical procedures (P<0.0001) on Bsx transcript levels. n=4, except for SCGdcn at ZT18, where n=3. Scale bar, 1 mm. B) Radiochemical in situ hybridization for detection of Bsx transcripts in the pineal gland of rats injected with phosphate buffered saline (PBS) or isoprenaline (ISO) during daytime. PBS, n=5; ISO, n=7. Scale bar, 1 mm. C) Bsx transcript levels in rat pineal cell cultures treated with norepinephrine (NE) or dibutyryl-cyclic adenosine monophosphate (DBcAMP) for 6h or 12h as determined by qRT-PCR. Two-way ANOVA detected significant effects of both stimulation type and duration (P<0.0001) on Bsx transcript levels. n=4. D) Bsx transcript levels in rat pineal cell cultures treated with actinomycin and/or stimulated with DBcAMP as determined by qRT-PCR. Two-way ANOVA detected significant effects of both actinomycin-treatment and DBcAMP-stimulation (P<0.0001) on Bsx transcript levels. n=3. Values on the graphs represent mean ± SEM. P-values determined by Welch’s or Student’s t-tests with Bonferroni corrections (A, C, D) or uncorrected Welch’s t-test (B): ns, not significant; *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001. ZT, Zeitgeber time.

To investigate if the Bsx rhythm in the pineal gland is driven by adrenergic signaling, rats were injected with isoprenaline during daytime, thus mimicking the nocturnal release of norepinephrine in the gland (Fig. 6B). Isoprenaline significantly increased Bsx mRNA levels (P<0.05), indicating that pineal Bsx expression is directly under the influence of adrenergic signaling (Fig. 6B).

To further delineate the regulatory roles of adrenergic stimulation and the second messenger cAMP, Bsx mRNA levels in pinealocyte cultures were assessed by use of qRT-PCR following stimulation with NE or DBcAMP (Fig. 6C). These treatments exerted significant effects on Bsx expression (P<0.0001); Bsx transcript levels were significantly increased in NE- and DBcAMP-treated cultures after both 6h (P<0.0001) and 12h of stimulation (NE, P<0.001; DBcAMP, P<0.01).

To clarify if cAMP-stimulation of Bsx expression occurs at the transcriptional level, the effect of DBcAMP on Bsx expression was analyzed in pinealocyte cultures pretreated with the transcriptional inhibitor actinomycin D (Fig. 6D). Actinomycin D significantly affected Bsx expression dynamics (P<0.0001) and the DBcAMP-induced increase of Bsx mRNA levels was not detected in actinomycin D-pretreated pinealocytes (P>0.05), indicating that upregulation of Bsx via cAMP requires de novo synthesis of Bsx mRNA.

Bsx is a negative regulator of the homeobox genes Pax4 and Otx2, but does not affect the expression of other important pineal homeobox genes or genes encoding melatonin-synthesizing enzymes

Since the pineal gland does not develop fully in Bsx knockouts¹⁴, we applied in vitro siRNA-mediated Bsx knockdown in pineal cell cultures to study the role of Bsx in the adult pineal gland. Bsx was knocked down in cultured cells treated with siRNA targeting Bsx (siBsx) and stimulated with NE to mimic the nocturnal situation in the gland (Fig. 7). The knockdown was verified by use of qRT-PCR: siBsx significantly reduced Bsx mRNA levels (P<0.0001) by approximately 80% (Fig. 7A).

Figure 7. The effects of siRNA-mediated Bsx knockdown on expression of genes encoding melatonin-synthesizing enzymes and homeodomain proteins in cultured pineal cells.

Analyses of Bsx (A), Tph1 (B), Aanat (C), Asmt (D), Pax4 (E), Rax (F), Otx2 (G), and Crx (H) transcript levels in rat pineal cell cultures treated with non-targeting siRNA (siNT) or siRNA targeting Bsx (siBsx) and stimulated with norepinephrine (NE) for 6h or 12h; transcript levels determined by qRT-PCR. The knockdown of Bsx by siBsx corresponded to 83%, 85%, and 71% reductions at 0h, 6h, and 12h, respectively. Two-way ANOVA detected significant effects of NE on transcript levels of Bsx (P<0.0001), Tph1 (P<0.05), Aanat (P<0.0001), Rax (P<0.0001), Crx (P<0.0001), Pax4 (P<0.0001), and Otx2 (P<0.01), and significant effects of siRNA on transcript levels of Bsx (P<0.0001), Pax4 (P<0.0001), and Otx2 (P<0.01). Values on the graphs represent mean ± SEM, n=6. P-values determined by Welch’s or Student’s t-tests with Bonferroni corrections: ns, not significant; **, P<0.01; ***, P<0.001; ****, P<0.0001.

To investigate the possible role of Bsx in regulation of melatonin synthesis, transcript levels of the genes encoding the enzymes responsible for this process were analyzed. As expected, NE stimulation increased the levels of Aanat (P<0.0001)³, but significant effects of siBsx on expression of Tph1, Aanat or Asmt at specific time-points after NE-stimulation were not detectable (P>0.05) (Fig. 7B–D).

The possible role of Bsx in regulation of other pineal homeobox genes was investigated by determining the effect of Bsx knockdown on expression of the paired box 4 (Pax4), retinal and anterior neural fold (Rax), orthodenticle 2 (Otx2), and cone-rod (Crx) homeobox genes (Fig. 7E–H). NE stimulation exerted a significant effect on all four homeobox genes (Otx2, P<0.01; Pax4, Rax, Crx, P<0.0001) in agreement with their rhythmic nature in vivo⁴. Differences in transcript levels between siNT- and siBsx-treated pinealocyte cultures were not detectable for Rax and Crx (P>0.05) (Fig. 7F,H). However, siBsx-treatment exerted significant effects on Pax4 (P<0.0001) with increased levels at all three time points (0h and 6h, P<0.001; 12h, P<0.0001), suggesting that Bsx is a negative regulator of Pax4 (Fig. 7E). Knockdown of Bsx affected Otx2 positively (P<0.01), indicating that Bsx may also negatively regulate Otx2, but significantly higher levels of Otx2 transcripts were only detected after 12 h of NE stimulation (P<0.0001) (Fig. 7G).

RNA sequencing identifies Bsx as a regulator of the adult rat pineal transcriptome

The global effects of Bsx on the rat pineal transcriptome were investigated using RNA sequencing of pineal cell cultures treated with siBsx or siNT and stimulated with DBcAMP for 6h to mimic nighttime. A PCA plot showed separate clustering for the two treatments (Supporting Information, Fig. S1). The siBsx-mediated knockdown of Bsx was confirmed (P<0.0001, 79% decrease) (Fig. 8). Minor differences, though significant with the statistical approach used here, were detected for the genes encoding key pineal homeodomain proteins and melatonin pathway enzymes (P-values < 0.05); however, only Pax4 and Rax showed changes in relative expression greater than 25% (P<0.01; 143% and 39% increase in siBsx-cultures, respectively) (Fig. 8). After filtering, a total of 801 differentially expressed genes, 372 upregulated and 427 downregulated in siBsx-treated cultures, were included in a DAVID analysis. Examining output for the GOterm Biological Processes at level 5 revealed a significant overrepresentation of genes involved in five types of developmental processes (Bonferroni-corrected P-values < 0.01), reflecting the role of Bsx as a developmental regulator (Table 3; Supporting Information, Table S1).

Figure 8. The effects of siRNA-mediated Bsx knockdown on expression of genes encoding melatonin-synthesizing enzymes and homeodomain proteins in cultured pineal cells as determined by RNA sequencing.

Changes in expression of genes encoding homeodomain proteins and melatonin-synthesizing enzymes in pineal cell cultures treated with siRNA targeting Bsx (siBsx) and stimulated with DBcAMP for 6 hours as determined by RNA sequencing. Values on the Y-axis are given as percentagewise relative expression compared to controls treated with non-targeting siRNA (siNT). P-values determined by Wald tests with Benjamini-Hochberg corrections: *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001.

Table 3. The effects of siRNA-mediated knockdown of Bsx on the rat pinealocyte transcriptome determined by RNA sequencing.

DAVID output for the GOterm Biological Processes at level 5 with significant overrepresentation in siBsx-treated pineal cell cultures compared to siNT-treated cultures as determined by RNA sequencing.

Gene function	% of filtered genes mapped to gene function	Significance level
Animal organ development	26.34%	****
Nervous system development	17.60%	**
Cell development	17.47%	***
Neuron development	10.43%	***
Sensory organ development	6.65%	**

Bonferroni-corrected P-values below 0.01 were considered statistically significant

^****, P<0.0001

^***, P<0.001

^**, P<0.01.

Discussion

Current knowledge establishes a role of Bsx in development of the pineal gland in several vertebrate species^13–16; in addition, genome-wide large-scale screening efforts suggest rhythmic expression of Bsx in the mature pineal gland of the rat¹⁷. The present study expands upon this knowledge by exploring persistent ontogenetic expression, cellular localization, regulation, and molecular role of Bsx in the rat pineal gland, as well as anatomical distribution of BSX transcripts in the human pineal gland.

Development of the pineal gland in mammals is guided by a number of homeobox gene-encoded transcription factors⁴. Bsx is required for proper development of the murine pineal gland along with Lhx9, Pax6, and Otx2^5,6,14,28, while other pineal homeobox genes, such as Rax, Crx, and Pax4^7–9, seem to exert their function at postnatal stages. Bsx and Otx2⁷ fall into both categories by being continuously expressed from development into adulthood in the rat pineal gland. The essential role for Bsx in rodent pineal gland development represents a challenge in loss-of-function experiments on the mature pineal gland¹⁴. To circumvent the hypoplastic pineal gland of the Bsx knockout model, we here utilized an siRNA-based method for in vitro gene silencing. Although this approach does not fully attenuate Bsx expression levels, it represents an advantage over knockout models by allowing tissue-specific knockdown in pinealocytes from a melatonin-proficient mature gland.

By employing siRNA-mediated knockdown in cultured rat pinealocytes, we have previously shown that Otx2 and Crx are required for obtaining high expression of all three genes encoding the melatonin-synthesizing enzymes, namely Tph1, Aanat, and Asmt¹², and studies in a conditional knockout mouse suggest that Rax controls expression of pineal Aanat¹¹. However, in the current study we saw no effect on mRNA levels of the enzymes in the melatonin synthesis pathway following siRNA-mediated knockdown of Bsx in pinealocyte cultures. Therefore, as opposed to other pineal homeobox genes, there is no evidence that Bsx play a role as a central regulator of melatonin production in the adult, despite prominent nocturnal expression in melatonin-proficient pinealocytes. However, this does not exclude the possibility that it might act at early stages of development to irreversibly influence the expression of one or more of these genes.

On the other hand, in a search for targets among other pineal homeobox genes expressed in adults, we found Bsx to be a negative regulator of Pax4. As previously reported, homeobox genes and Aanat exhibit circadian expression profiles with a distinct sequential pattern in times of peak mRNA levels in the mature pineal gland, in the order of Pax4 (mid-to-late day), Rax (day-night transition), Otx2 (early-to-mid night), Crx (mid night), and Aanat (mid-to-late night)⁴. The role of Bsx as a negative regulator of Pax4 presented in this study fits into this model, as the rhythm of Bsx expression with a peak in the middle of the night is anti-phasic to that of Pax4, which peaks during the day. Similarly, previous studies have described mutual regulation of Crx and Otx2, as knockdown of Otx2 decreased Crx mRNA levels, while knockdown of Crx resulted in increased mRNA levels of Otx2^5,12. In our RNAseq analyses, Bsx knockdown influenced a number of pineal homeobox genes, while, in addition to Pax4, qPCR analyses only confirmed an effect on Otx2 after 12h of NE stimulation. These apparent discrepancies probably reflect different statistical approaches and cutoff criteria. The fact that homeobox genes, including both Bsx itself and its downstream targets, would normally guide development of various cell types is reflected in the results of the DAVID analysis of our RNAseq data indicating that Bsx influences a number of developmental processes. However, in addition to regulation of developmental processes and melatonin synthesis, our data suggest the presence of an internal regulatory interplay between pineal homeobox genes, which might also explain the differences in circadian expression profiles despite the genes being under control of the same adrenergic input to the gland.

Our data show that Bsx expression is circadian in nature and driven by norepinephrine released from sympathetic nerve endings acting via cAMP as a second messenger to induce nocturnal transcription of Bsx. The same signaling mechanism drives the nocturnal increase in Aanat mRNA levels and other transcripts in the rat pineal transcriptome^1,17,29, including pineal homeobox genes as described above. However, unlike other pineal homeobox genes governed by the same regulatory mechanisms, Bsx expression was not detected in the retina. Otx2, Crx, Rax, Pax6, and Pax4 are highly expressed in both tissues at various developmental stages in the rat^9,30,31, reflecting the close evolutionary relationship between the pinealocyte and the retinal photoreceptor cell³². Bsx may be a determining factor required for obtaining and maintaining the pinealocyte phenotype; the underlying transcriptional network may otherwise be similar to that of the retinal photoreceptor. This would imply mammalian Bsx to be a potential determining factor for pineal cell fate during development, as has been suggested in studies on non-mammalian vertebrates, including Xenopus and zebrafish^15,16,33. In contrast to mammals, the pineal gland of non-mammalian vertebrates is a photoreceptive organ with pineal photoreceptors sharing common ancestry with the mammalian pinealocyte^34,35. In both Xenopus and zebrafish, Bsx homologs were shown to control specification of photoreceptive pinealocytes^15,16,33,35. Expression of BSX in the human pineal gland, as reported here, is consistent with a view of a conserved role for pineal Bsx across species.

Abbreviations

Aanat

arylalkylamine N-acetyltransferase

Actb

actin beta

Asmt

acetylserotonin O-methyltransferase

Bsx

brain-specific homeobox gene

Crx

cone-rod homeobox

CT

circadian time

DBcAMP

dibutyryl cyclic adenosine monophosphate

DD

constant darkness

E

embryonic day

12L:12D

12 h light:12 h dark schedule

Gapdh

glyceraldehyde 3-phosphate dehydrogenase

NE

norepinephrine

Otx2

orthodenticle homeobox 2

P

postnatal day

Pax4

paired box 4

PBS

phosphate buffered saline

Rax

retinal and anterior neural fold homeobox

siRNA

short interfering RNA

SCGdcn

decentralization of the superior cervical ganglia

SCGx

superior cervical ganglionectomy

Tph1

tryptophan hydroxylase 1

ZT

Zeitgeber time