A chromatin-based mechanism controls differential regulation of the cytochrome P450 gene Cyp24a1 in renal and non-renal tissues

Abstract

Cytochrome P450 family 27 subfamily B member 1 (CYP27B1) and CYP24A1 function to maintain physiological levels of 1,25-dihydroxyvitamin D₃ (1,25(OH)₂D₃) in the kidney. Renal Cyp27b1 and Cyp24a1 expression levels are transcriptionally regulated in a highly reciprocal manner by parathyroid hormone (PTH), fibroblast growth factor 23 (FGF23), and 1,25(OH)₂D₃. In contrast, Cyp24a1 regulation in nonrenal target cells (NRTCs) is limited to induction by 1,25(OH)₂D₃. Herein, we used ChIP-Seq analyses of mouse tissues to identify regulatory regions within the Cyp24a1 gene locus. We found an extended region downstream of Cyp24a1 containing a cluster of sites, termed C24-DS1, binding PTH-sensitive cAMP-responsive element–binding protein (CREB) and a cluster termed C24-DS2 binding the vitamin D receptor (VDR). VDR-occupied sites were present in both the kidney and NRTCs, but pCREB sites were occupied only in the kidney. We deleted each segment in the mouse and observed that although the overt phenotypes of both cluster deletions were unremarkable, RNA analysis in the C24-DS1–deleted strain revealed a loss of basal renal Cyp24a1 expression, total resistance to FGF23 and PTH regulation, and secondary suppression of renal Cyp27b1; 1,25(OH)₂D₃ induction remained unaffected in all tissues. In contrast, loss of the VDR cluster in the C24-DS2–deleted strain did not affect 1,25(OH)₂D₃ induction of renal Cyp24a1 expression yet reduced but did not eliminate Cyp24a1 responses in NRTCs. We conclude that a chromatin-based mechanism differentially regulates Cyp24a1 in the kidney and NRTCs and is essential for the specific functions of Cyp24a1 in these two tissue types.

Introduction

Vitamin D undergoes final bioactivation in the kidney via Cyp27b1-mediated 1α-hydroxylation to 1,25-dihydroxyvitamin D₃ (1,25(OH)₂D₃),³ the hormonal form of the vitamin (1). Whereas 1,25(OH)₂D₃ arises exclusively from Cyp27b1 activity, the actual blood levels of 1,25(OH)₂D₃ are also strongly influenced by the catabolic activity of renal CYP24A1 (2, 3). This second enzyme is active in higher concentrations than CYP27B1 in kidney mitochondria and is responsible for the degradation of both 25(OH)D₃ and 1,25(OH)₂D₃ to the initial intermediates 24,25(OH)₂D₃ and 1,24,25(OH)₃D₃, respectively. The same enzyme then performs further multistep catabolism of each of these products to specific acids, which leave the body through bile (4, 5). Whereas C24-hydroxylation is the preferred catabolic route in humans and rodents, CYP24A1 can also C23-hydroxylate 25(OH)D₃ and 1,25(OH)₂D₃, which culminate in terminal 26,23-lactone products; both pathways predominate in the opossum and guinea pig. New metabolite profiling methods based on LC-tandem MS are now able to detect a range of serum metabolites formed by CYP24A1, including 24,25(OH)₂D₃, 1,24,25(OH)₃D₃, and 25(OH)D₃-26,23-lactone. It is questionable whether any of these metabolites are biologically active, although recent studies by St-Arnaud and colleagues (6) have suggested that 24,25(OH)₂D₃ may act as an allosteric activator of FAM57B2 that is involved in the synthesis of lactosylceramide, providing a potential mechanism for an early hypothesized role for 24,25(OH)₂D₃ in bone fracture healing. 1,24,25(OH)₃D₃ is also of particular interest because recent estimates of its circulating level in mice suggest that this metabolite might contribute to the net biological activity of vitamin D₃ (7–10).

In nonrenal vitamin D target cells, inactivation of 1,25(OH)₂D₃ by CYP24A1 likely predominates over that of 25(OH)D₃; thus, the degradation of endocrine 1,25(OH)₂D₃ as well as hormone potentially produced in these cells via local Cyp27b1 expression may impact 1,25(OH)₂D₃'s ability to control cell-specific biological actions via set point modification. The renal function of CYP24A1, however, is highlighted in the Cyp24a1-null mouse, wherein the loss of this enzyme's actions leads to an initial elevation in the circulating levels of 25(OH)D₃ and 1,25(OH)₂D₃, although secondary homeostatic events appear to down-regulate Cyp27b1 expression in response to the reduction in 1,25(OH)₂D₃ turnover (11, 12). The critical role of Cyp24a1 was strongly reinforced more recently in humans with idiopathic infantile hypercalcemia, where, due to mutation in the gene, a defective CYP24A1 protein fails to degrade 1,25(OH)₂D₃, leading to hypercalcemia and kidney disease (13). These patients exhibit significantly reduced serum 24,25(OH)₂D₃ and elevated 25(OH)D₃/24,25(OH)₂D₃ ratios. This observation has been recapitulated in the Cyp24a1-null mouse, confirming that serum vitamin D metabolite profiles serve as an effective biomarker of Cyp24a1 expression and enzymatic activity in vivo.

The Cyp24a1 gene is regulated in the kidney by the same hormones that control Cyp27b1 expression, albeit in a reciprocal fashion (1). Thus, whereas PTH induces Cyp27b1 expression, it strongly suppresses Cyp24a1; FGF23 and 1,25(OH)₂D₃, on the other hand, strongly induce renal Cyp24a1 yet suppress renal Cyp27b1 (7, 14, 15). This regulatory paradigm serves both to enhance the production of 1,25(OH)₂D₃ and to suppress its degradation under conditions of reduced 1,25(OH)₂D₃ synthesis, thereby raising circulating levels of the hormone. High levels of 1,25(OH)₂D₃, in contrast, decrease PTH and raise FGF23 levels, collectively suppressing CYP27B1-mediated production of 1,25(OH)₂D₃ while increasing CYP24A1-directed degradation. Thus, reciprocal regulation of Cyp27b1 and Cyp24a1 in the kidney coordinately functions to raise or lower 1,25(OH)₂D₃ levels to maintain physiologically appropriate concentrations of the hormone and thus to sustain normal extracellular mineral homeostasis (16). This reciprocal regulation does not occur in nonrenal target cells (NRTCs), however, where Cyp24a1 is solely up-regulated by 1,25(OH)₂D₃, potentially distancing the actions of the mineral-regulating hormones PTH and FGF23 on Cyp24a1 expression in the kidney from those that regulate its actions in peripheral target tissues (7). Accordingly, whereas CYP27B1 is necessary for 1,25(OH)₂D₃ production, CYP24A1 contributes to the adaptive vitamin D metabolism that controls blood 1,25(OH)₂D₃ levels relevant to the maintenance of mineral homeostasis.

Early studies of the Cyp24a1 gene revealed the presence of two vitamin D response elements active in mediating the response to 1,25(OH)₂D₃ (17–22). This activity was present in virtually all cell types, including the kidney. More recently, however, we have shown using unbiased ChIP-chip and ChIP-Seq analyses that the regulation of Cyp24a1 by 1,25(OH)₂D₃ is also mediated by a cluster of intergenic components located downstream of the gene in both mouse and human nonrenal cells that contain functionally active vitamin D response elements, which bind the VDR upon 1,25(OH)₂D₃ activation (23). Mechanisms that mediate both PTH and FGF23 actions in the kidney are unclear, however (7, 24–26). Finally, it is also worth noting that whereas 1,25(OH)₂D₃, PTH, and FGF23 are primary regulators, other hormones, cytokines, and systemic components as well as intracellular chromatin coregulatory factors also modulate Cyp24a1 expression (27–32).

Our previously reported studies revealed a unique regulatory module in the mouse kidney that is essential for the novel regulation of Cyp27b1 by PTH, FGF23, and 1,25(OH)₂D₃ in vivo (7, 26). Given the reciprocal nature of Cyp24a1 regulation in the kidney by these same hormones together with the general absence of regulation by PTH and FGF23 in NRTCs, we initiated studies herein aimed at illuminating the underlying genomic basis for this differential regulation using both ChIP-Seq analyses and CRISPR/Cas9 gene editing methods. Our results identify a genomic mechanism in mice through which the differential expression of Cyp24a1 occurs in the kidney and in NRTCs and highlight the role of this regulation as a primary determinant of the separate functions of Cyp24a1 in vitamin D metabolism.

Results

Differential regulation of Cyp24a1 by PTH, FGF23, and 1,25(OH)₂D₃ in the kidney and in NRTCs

The results in Fig. 1A demonstrate that Cyp24a1 expression in the kidney is up-regulated in response to a single time- and dose-optimized injection of FGF23 or 1,25(OH)₂D₃ in the mouse and suppressed by a similar injection of PTH, a regulatory pattern that is the reciprocal of that seen for Cyp27b1 in the same tissue (7, 26). This control in NRTCs is limited to up-regulation by 1,25(OH)₂D₃, however, demonstrating a collective differential regulatory response to these hormones in nonrenal tissues. It is also worth noting that the basal expression of Cyp24a1 in the kidney is much higher than that seen in NRTCs, an observation that is now well-established. Indeed, in some NRTC tissues, basal expression levels are almost undetectable (15). The underlying determinants of elevated Cyp24a1 expression in the kidney are not fully understood, although it is likely that the circulating levels of PTH, FGF23, and perhaps 1,25(OH)₂D₃ contribute to this basal expression (15). Regardless of this, the differential regulation of Cyp24a1 with respect to PTH and FGF23 in NRTCs is unique. Because some of these tissues contain mediators for both PTH and FGF23 action, it seemed unlikely that the lack of response was due to the absence of appropriate signaling pathways.

Figure 1.

Transcriptional regulation of Cyp24a1 in kidney and nonrenal target cells. A, gene expression of Cyp24a1 after treatment in 8–9-week-old C57BL/6 WT mice with ethanol/PBS vehicle (Veh, gray, n = 6), 230 ng/g bw PTH for 1 h (PTH, blue, n = 6), 10 ng/g bw 1,25(OH)₂D₃ (1,25D₃, black, n = 6) for 6 h, or 50 ng/g bw FGF23 for 3 h (FGF23, green, n = 6) examined in kidney, intestine, skin, calvaria, L5 vertebrae, and TPTG. Data are displayed as relative quantitation (RQ, mean ± S.E.) compared with Gapdh. *, p < 0.05 paired t test: treatment versus vehicle. B, overlaid ChIP-Seq data tracks for H3K36me3 from mouse kidney displayed as either basal or vehicle (yellow, n = 3) or treated (blue, n = 3) with 1,25(OH)₂D₃, FGF23, or PTH as indicated with concentrations used in A with a treatment time of 1 h. Overlapping track data appear as green. Regions of interest are highlighted in light gray boxes. Genomic location and scale are indicated (top), and maximum height of tag sequence density for each data track is indicated on the y axis (top left of each track, normalized to input and 10⁷ tags). Gene transcriptional direction is indicated by an arrow, and exons are indicated by boxes.

PTH, FGF23, and 1,25(OH)₂D₃ regulate Cyp24a1 expression via transcription

Although the actions of 1,25(OH)₂D₃ on Cyp24a1 expression as seen in Fig. 1A are known to be transcriptional in nature (7, 21), the underlying mechanisms responsible for PTH suppression and FGF23 induction have been unclear. Thus, for example, it has been suggested that PTH suppression of Cyp24a1 expression is mediated via a post-transcriptional mechanism, whereas modulation of the gene by FGF23 is entirely unknown (25, 33). We therefore conducted a ChIP-Seq analysis of the mouse genome, focusing entirely on the ability of the three hormones to alter the content of the histone mark H3K36me3 directly at the Cyp24a1 locus. This mark has generally been linked mechanistically to levels of transcriptional elongation across most genetic loci (34). As can be seen in Fig. 1B, residual H3K36me3 enrichment was present exclusively within the Cyp24a1 transcription unit in the kidney of WT mice, as predicted for this chromatin modification. Importantly, however, a 1-h treatment of each of the three hormones resulted in a striking increase in the histone mark in response to 1,25(OH)₂D₃ and FGF23 (blue) and a modest suppression in response to PTH (yellow). Although these data do not prove that PTH and FGF23 operate at the transcriptional level, they certainly support the idea that, like 1,25(OH)₂D₃, these hormones act directly to influence the selective output of Cyp24a1 transcripts. As can also be seen in Fig. 1B, no changes in H3K36me3 density were observed at adjacent Pfdn4, where the transcript for this gene in the kidney is generated, or downstream of the gene at Bcas1, which is not transcribed (data not shown).

ChIP-Seq analysis reveals sites of VDR and pCREB binding at the Cyp24a1 gene locus in the kidney that are restricted in NRTCs to those that bind the VDR

Our preliminary examination of the CYP24A1 gene in several mouse and human nonrenal cell lines using ChIP-Seq analysis suggested the 1,25(OH)₂D₃-inducible presence of the VDR at novel sites located in the intergenic region downstream of CYP24A1 as well as immediately upstream of the Cyp24a1 promoter (23, 35, 36). Using this same technique in mouse tissues, the results in Fig. 2A confirm that 1 h after 1,25(OH)₂D₃ injection, the VDR was present at various densities in these same downstream (termed herein Cyp24a1-downstream 2 (C24-DS2)) and upstream (termed Cyp24a1 promoter–proximal 1 (C24-PP1)) sites in isolated nonrenal mesenchymal cells, small intestinal tissue, thyroparathyroid glands (TPTG), and bone marrow cells and also in kidney tissue. Importantly, ChIP-Seq analysis of the kidney also revealed the modest presence of phosphorylated CREB (pCREB), a major protein kinase A/PTH transcription factor target, at several sites in C24-DS2 but most strongly at three additional novel sites in a region (termed Cyp24a1-downstream 1 (C24-DS1)) located more proximal to the 3′ end of the Cyp24a1 gene yet outside the VDR binding cluster itself. pCREB was apparent in the C24-PP1 region of Cyp24a1 as well. As noted earlier, Pfdn4 was not regulated by any of the three hormones here or in NRTCs and is therefore not a target gene. A similar analysis 1 h after PTH injection revealed a modest effect to reduce pCREB (yellow) levels within C24-DS1 that could represent a mechanistic determinant of PTH suppression, although additional studies will be required to firmly establish this finding. Interestingly, as documented in Fig. 2B, neither residual nor inducible pCREB was present across the entire region of C24-DS1 or in C24-PP1 in mesenchymal stem cells (MSC) or peripheral blood monocytes (pBMCs) or in any other NRTCs we examined (data not shown). These results suggest that whereas VDR is capable of occupying sites and mediating activity at the Cyp24a1 gene in both the kidney and in representative NRTCs, the absence of pCREB binding and perhaps other factor activities likely restrict the effects of PTH and FGF23 in NRTCs.

Figure 2.

ChIP-Seq analysis displays kidney-specific regulation at the Cyp24a1 gene locus. A–D, overlaid ChIP-Seq data tracks for VDR (A), pCREB (B), H3K4me1 (C), and H3K9ac (D) from mouse kidney, bone marrow–derived mesenchymal stem cells (MSC), intestine, TPTG, or peripheral blood monocytes (pBMC) displayed as either basal or vehicle (yellow, n = 3) or treated (blue, n = 3) with PTH, forskolin (FSK), 1,25(OH)₂D₃, or FGF23, as indicated. E, DNase-Seq data from the ENCODE project for kidney and intestine. F, H3K9ac data in VDRKO (blue), Cyp27b1KO mice (C27B1KO, blue), or M1-IKO (blue) versus WT (yellow). Overlapping track data appear as green. Regions of interest are highlighted in light gray boxes denoted as C24-DS1, C24-DS2, and C24-PP1. Genomic location and scale are indicated (top), and maximum height of tag sequence density for each data track is indicated on the y axis (top left of each track, normalized to input and 10⁷ tags). Gene transcriptional direction is indicated by an arrow, and exons are indicated by boxes.

H3K4me1 and H3K9ac enrichment co-localizes with both VDR- and pCREB-occupied sites at the Cyp24a1 gene locus in the kidney but is selectively absent at unoccupied pCREB sites in the intestine

To assess whether sites of VDR- and pCREB-binding activity were contained within active enhancers, we conducted a further ChIP-Seq analysis of H3K4me1 and of H3K9ac across the Cyp24a1 gene locus; the former histone modification represents a classic epigenetic enhancer signature mark, whereas the latter represents an epigenetic mark that is frequently indicative of altered enhancer activity (34, 37). Accordingly, mice were treated with either vehicle, PTH, FGF23, or 1,25(OH)₂D₃ for 1 h and then subjected to ChIP-Seq analysis as in the earlier studies of H3K36me3. As seen in Fig. 2C, basal enrichment of the H3K4me1 mark was evident not only at C24-PP1 but also at a number of broad sites that align directly with sites that retain VDR (C24-DS2) and/or pCREB (C24-DS1) across the intergenic segment located downstream of the Cyp24a1 locus. Both 1,25(OH)₂D₃ and FGF23 modestly induce H3K4me1 densities at each of these potential enhancer regions, although the suppressor PTH is without effect on this mark. Importantly, whereas significant H3K4me1 enrichment aligned within the VDR-binding regions in the intestine, no such enrichment was evident within the C24-DS1 segment that binds pCREB exclusively in the kidney. As seen in Fig. 2D, similar basal and regulatory profiles were observed for the H3K9ac enhancer activity mark. In this case, however, whereas 1,25(OH)₂D₃ actively induces the H3K9ac mark, FGF23 had at best a modest effect on this modification (1,25(OH)₂D₃ is a much stronger inducer), whereas PTH was seen to suppress this mark strongly, demonstrating expected selectivity relative to H3K9ac. Again, however, whereas the H3K9ac mark was present and aligned with VDR occupancy and H3K4me1 enrichment in the intestine (C24-DS2) and was also seen to increase in response to 1,25(OH)₂D₃, no such activity was evident across the C24-DS1 segment unable to bind pCREB. This was also true for other NRTC tissues (data not shown).

This extended downstream region also contained open chromatin sites, as identified through studies conducted by the ENCODE Consortium in both kidney and intestine. In these experiments, DHS analysis of the tissues of 8-week-old mice revealed the presence of well-defined open chromatin sites that aligned directly with both VDR- and pCREB-bound enhancers (C24-DS2 and C24-DS1, respectively) in the kidney but were restricted exclusively to VDR C24-DS2 sites in the intestine (Fig. 2E). Finally, we conducted ChIP-Seq analysis of H3K9ac in the kidneys of Vdr- and Cyp27b1-null mice, which have highly suppressed Cyp24a1 expression levels due to highly elevated PTH and strongly suppressed FGF23 concentrations in the blood (7, 15, 26), and in M1-IKO mice where elevated PTH and lowered FGF23 parameters also exist, although for different mechanistic reasons (see Table 1) (7, 26). As can be seen in Fig. 2F, whereas H3K9ac enrichment was present in WT mice, there was no evidence of H3K9ac activity at sites across the entire Cyp24a1 gene locus in any of these mouse strains. These data not only support the relevance of the H3K9ac mark for Cyp24a1 activity but also demonstrate that despite the presence (M1-IKO mice) or absence (Cyp27b1-null mice) of 1,25(OH)₂D₃ or the absence of 1,25(OH)₂D₃ activity (Vdr-null mice), high PTH and low FGF23 levels prevent the expression and regulation of Cyp24a1. In addition, these overall observations indicate that the factor that mediates FGF23 activity is likely to bind to the same region to which pCREB alone is bound.

Table 1.

Animal models employed

Strain/type	Phenotype	Reference
WT	None
C24-DS1KO	Low Cyp24a1 basal, reciprocally low Cyp27b1 basal	This work
C24-DS2KO	Cyp24a1 1,25(OH)2D3-sensitive in NRTC only	This work
Cyp27b1-null	Cyp27b1-null, low Cyp24a1 basal	Refs. 7 and 26
Cyp24a1-null	Cyp24a1-null	Ref. 11
M1-IKO	Low Cyp27b1 basal, Cyp27b1 PTH-insensitive	Refs. 7 and 26
M21-IKO	Low Cyp27b1 basal, Cyp27b1 FGF23-insensitive	Refs. 7 and 26
M1/M21-DIKO	Low Cyp27b1 basal, Cyp27b1 PTH/FGF23/1,25(OH)2D3-insensitive	Ref. 26

Establishing the functional role of the DS1 regulatory region

The large downstream regulatory domain for Cyp24a1 that contains multiple components that generally favor either VDR (C24-DS2) or pCREB (C24-DS1) binding is clearly complex in its linear arrangement, spanning almost 30 kb of DNA. In addition, whereas the distinction between the numerous VDR and pCREB binding sites is clear, particularly in the kidney versus other NRTCs, it is also evident that modest transcription factor overlap may exist at a subset of these sites in the kidney. Given this complexity, we employed a CRISPR/Cas9 gene editing approach as documented in Fig. 3 and deleted either 17.5 kb of DNA containing the three dominant pCREB binding sites or 2.4 kb containing the two major sites of VDR binding and created two mutant mouse strains designated Cyp24-DS1KO (C24-DS1KO, 17.5-kb deletion) and Cyp24-DS2KO (C24-DS2KO, 2.4-kb deletion), respectively (7). Male and female C24-DS1KO mice appeared normal in size, weight, and appearance as compared with WT littermate controls, and their skeletons exhibited normal whole-body, femur, and spine bone mineral densities (BMDs), as documented in Table 2. Interestingly, whereas both blood calcium (Ca) and phosphate (P) levels were also comparable with controls, providing the basis for the normal skeletal BMDs that were observed, PTH levels were modestly suppressed, whereas FGF23 levels were somewhat elevated (Table 3). These results suggested that although mineral homeostasis was normal, minor hormonal adaptations might accommodate the maintenance of Ca and P levels within the normal range. Based upon these initial observations, we isolated kidney RNA and examined the overall expression of Cyp24a1. As can be seen in Fig. 4 (A–C), basal Cyp24a1 expression in vehicle-treated mice was dramatically reduced in the kidney, whereas Cyp24a1 response to a single injection of either FGF23 (Fig. 4A) or PTH (Fig. 4B) was fully abrogated relative to similarly treated littermate controls. These results support the hypothesis that this large regulatory segment in normal mice likely mediates both PTH and FGF23 modulation. Interestingly, the noted reduction in Cyp24a1 expression in the kidneys of C24-DS1KO mice was also accompanied by a secondary reduction in Cyp27b1 expression, as seen in Fig. 4A, likely the result of homeostatic adaptation prompted by decreased PTH and increased FGF23 levels, as discussed previously (Table 2). This coordinated secondary reduction in Cyp27b1 is analogous to that seen for Cyp24a1 when Cyp27b1 expression is reduced in Cyp27b1-null, M1-IKO, and M1/M21-DIKO mice as a result of elevated PTH and reduced FGF23 levels (see Table 4) (7, 26) and appears to be a hallmark of the reciprocal regulation of these two genes.

Figure 3.

Schematic representation of the CRISPR/Cas9 deletions in the downstream regions of the Cyp24a1 locus. VDR and pCREB ChIP-Seq data (as in Fig. 2) are shown for reference. Highlighted regions were excised using two CRISPR-guide RNAs each for Cyp24a1-DS1KO and Cyp24a1-DS2KO. Both regions shared a common guide RNA (G2).

Table 2.

Bone mineral densities and body weights (n = 6–13)

	Sex	WT	DS1
Whole body (g/cm2)	Female	0.045 ± 0.001	0.045 ± 0.001
	Male	0.049 ± 0.001	0.049 ± 0.001
Femur (g/cm2)	Female	0.060 ± 0.001	0.060 ± 0.001
	Male	0.074 ± 0.001	0.074 ± 0.001
Spine (g/cm2)	Female	0.058 ± 0.002	0.056 ± 0.001
	Male	0.062 ± 0.001	0.061 ± 0.001
Body weight (g)	Female	19.8 ± 0.5	19.5 ± 0.2
	Male	24.6 ± 0.3	24.4 ± 0.4

Table 3.

Systemic factor analyses (n = 13)

*, p < 0.05 versus WT littermates.

Factor	WT	DS1
Calcium (mg/dl)	10.7 ± 0.2	10.9 ± 0.2
Phosphate (mm)	2.7 ± 0.1	2.6 ± 0.1
PTH (pg/ml)	48.2 ± 1.4	39.8 ± 0.3*
FGF23 (pg/ml)	244.4 ± 10.2	319.7 ± 13.0*

Figure 4.

Transcriptional regulation of Cyp24a1 in Cyp24-DS1KO mice. A–C, expression of the genes indicated after treatment in 8–9-week-old Cyp24-DS1KO mice (DS1) and their WT littermates (WT) with ethanol/PBS vehicle (Veh, gray, n = 6–10), 50 ng/g bw FGF23 for 3 h (FGF23, green, n = 6), 230 ng/g bw PTH for 1 h (PTH, blue, n = 8), or 10 ng/g bw 1,25(OH)₂D₃ (1,25D₃, black, n = 9) for 6 h was examined in kidney. D, gene expression of Cyp24a1 after treatment in 8–9-week-old DS1 mice (DS1) and their WT littermates (WT) with ethanol vehicle (Veh, gray, n = 8–9), or 10 ng/g bw 1,25(OH)₂D₃ (1,25D₃, black, n = 9) for 6 h was examined in nonrenal tissues as indicated. Data are displayed as relative quantitation (RQ, mean ± S.E. (error bars)) compared with Gapdh. *, p < 0.05, paired t test: treatment versus vehicle. #, p < 0.05, paired t test: DS1 versus WT received the same treatment.

Table 4.

Vitamin D₃ metabolite analyses (n = 6–8)

ND, not determined; *, p < 0.05 versus WT littermates.

Metabolite	WT	DS1	WT	DS2
25(OH)D3 (ng/ml)	14.8 ± 1.0	51.0 ± 3.4*	16.8 ± 0.3	15.3 ± 0.9
24,25(OH)2D3 (ng/ml)	7.8 ± 0.8	26.5 ± 4.5*	8.1 ± 0.7	7.0 ± 0.3
25(OH)D3/24,25(OH)2D3 ratio	1.9 ± 0.3	2.0 ± 0.5	2.2 ± 0.4	2.2 ± 0.2
25(OH)D3-26,23-lactone (ng/ml)	3.1 ± 0.2	1.1 ± 0.1*	4.8 ± 0.2	4.4 ± 0.3
25(OH)D3/25(OH)D3-26,23-lactone ratio	4.9 ± 0.9	47.3 ± 12.4*	3.5 ± 0.5	3.5 ± 0.4
1,25(OH)2D3 (pg/ml)	24.6 ± 2.4	26.8 ± 1.4	21.2 ± 2.2	18.8 ± 1.3
1,24,25(OH)3D3 (pg/ml)	70.9 ± 2.9	44.8 ± 3.6*	68.8 ± 3.2	66.8 ± 4.2

Importantly, treatment of the mutant mice with 1,25(OH)₂D₃ resulted in a strong up-regulation of Cyp24a1 expression in the kidney (Fig. 4C), although this up-regulation was slightly compromised relative to WT mice. It is worth noting, however, that WT -fold induction by 1,25(OH)₂D₃ was ∼46-fold above baseline, whereas in the C24-DS1 mice, it was 740-fold. Because the basal expression of Cyp24a1 in C24-DS1KO mice is strongly reduced, it is difficult to ascertain at this level of resolution whether the decreased net response to 1,25(OH)₂D₃ is due to the absence of potentially overlapping sites of weak VDR binding in C24-DS1 or to the absence of regulatory sites responsible for basal Cyp24a1 expression in C24-DS1 that might influence the activity of 1,25(OH)₂D₃. It seems unlikely, however, that the reduced response to 1,25(OH)₂D₃ is due to changes in VDR DNA-binding activity present in the C24-DS2 segment, because Vdr concentrations at the level of RNA remain unaltered.

In contrast, examination of Cyp24a1 expression in multiple NRTCs, including intestine, bone, spleen, thymus, and skin, in the C24-DS1KO mouse (Fig. 4D) revealed that neither basal nor 1,25(OH)₂D₃ induction was compromised in any of these tissues with the exception of the TPTG, which, like the kidney, exhibited a reduced response to the hormone. Finally, it is interesting to note with respect to C24-DS1KO mice that because FGF23 induces whereas PTH suppresses Cyp24a1, the striking loss of basal expression of Cyp24a1 in these mice may be due entirely to the loss of responsivity to FGF23 induction. Because PTH is a suppressor, it is clear that loss of this response would result in derepression, an effect opposite that of FGF23. This interpretation would point to FGF23 as the dominant physiological regulator of basal Cyp24a1 expression in the kidney.

Vitamin D metabolism in C24-DS1KO mice

Based upon the systemic phenotype associated with the C24-DS1 mutation and the features of Cyp24a1 expression in the kidney and in NRTCs, we assessed in an additional examination whether the reduction in both Cyp24a1 and Cyp27b1 resulted in an aberrant level of 1,25(OH)₂D₃ and/or other vitamin D metabolites that comprise a vitamin D metabolite panel. As can be seen in Table 4, circulating levels of the 1,25(OH)₂D₃ hormone were normal in C24-DS1KO mice as compared with WT littermate controls and were consistent with normal mineral levels and skeletal BMDs seen in these mice as well. Interestingly, however, a significant alteration in the level of key metabolites of vitamin D was evident in the blood of C24-DS1KO mice that clearly resulted from the reduction in Cyp24a1 expression. Thus, in contrast to normal mice, where levels of 25(OH)D₃ and 24,25(OH)₂D₃ were maintained at a ratio of between 1.6 and 2.0, there was a 3–4-fold increase in the absolute amounts of both 25(OH)D₃ and 24,25(OH)₂D₃ coincident with a largely unchanged ratio between the two metabolites in C24-DS1KO mice. These findings support a markedly reduced catabolism of the 25(OH)D₃ substrate and a corresponding steady-state accumulation of the initial 24,25(OH)₂D₃ catabolic product of CYP24A1 activity. The decrease in CYP24A1 activity toward 25(OH)D₃ was particularly reinforced by the observed reduction in 25(OH)D₃-26,23-lactone, the metabolite arising from CYP24A1 catabolism via the 23,25(OH)₂D₃ metabolic pathway, and by the dramatic increase in the ratio of 25(OH)D₃/25(OH)D₃-26,23-lactone from 5 in WT mice to 47 in the C24-DS1 mouse, both consistent with reduced Cyp24a1 expression. Clearly, while the ratio of 25(OH)D₃ to the initial catabolic product of the 24,25(OH)₂D₃ degradation pathway (24,25(OH)₂D₃) remained largely unchanged, the ratio of 25(OH)D₃ to the terminal catabolic product of the 23,25(OH)₂D₃ degradation pathway was strikingly increased. Finally, C24-DS1KO mice also displayed lowered 1,24,25(OH)₃D₃ levels as compared with WT littermate controls, confirming impaired catabolism of the 1,25(OH)₂D₃ metabolites as well. This metabolic profile is consistent with that seen in other mouse models where Cyp24a1 levels are significantly reduced (see Table 1) (7, 26). We conclude that this mutant mouse with compromised Cyp24a1 and reduced Cyp27b1 expression levels exclusively in the kidney is able to maintain normal Ca and P homeostasis via a minor suppression in PTH and a modest elevation in FGF23 that significantly alters vitamin D metabolism to accommodate the maintenance of appropriate levels of 1,25(OH)₂D₃.

The role of Cyp24a1 in the regulation of circulating 25(OH)D₃

The vitamin D metabolic profile observed in the C24-DS1KO mouse above as well as profiles of a similar nature that had emerged in earlier mutant mouse models of altered Cyp24a1 expression suggested the possibility that in addition to its regulation of 1,25(OH)₂D₃, Cyp24a1 might be an important determinant of circulating 25(OH)D₃ levels as well. To explore this idea, we plotted the blood levels of 25(OH)D₃, 24,25(OH)₂D₃, and 25(OH)D₃-26,23-lactone as a function of basal renal Cyp24a1 expression in tissues obtained from WT mice (defined as 100%) and from several of the mutant mouse strains we alluded to above that exhibited variable decreases in basal Cyp24a1 expression levels (Cyp24a1 expression was indicated as a percentage of the level measured in WT littermate controls). The mutant strains utilized were M1-IKO, M21-IKO, and M1/M21-DIKO mice, which exhibit resistance to PTH and/or FGF23 or 1,25(OH)₂D₃; Cyp27b1-null mice, which cannot synthesize 1,25(OH)₂D₃; and Cyp24a1-null mice, which cannot catabolize either 25(OH)D₃ or 1,25(OH)₂D₃ (phenotypes summarized in Table 1) (7, 26, 38). We used these strains specifically because, with the exception of the Cyp24a1-null mouse, they each contained mutations at the Cyp27b1 locus that altered Cyp27b1 expression and 1,25(OH)₂D₃ production, which allowed for a common differential, homeostatic suppression of renal Cyp24a1 expression. As can be seen in Fig. 5, declining Cyp24a1 expression over a significant range beginning with WT levels correlated directly with modestly increasing concentrations of both 25(OH)D₃ and the initial metabolic degradation product 24,25(OH)₂D₃ such that the ratio of the two metabolites remained relatively constant between 1.6 and 2.0, ratios similar to that observed in the C24-DS1KO mouse (Table 4). An inflection point for Cyp24a1 expression was observed in M21-IKO mice, however, where 24,25(OH)₂D₃ levels began to decrease while 25(OH)D₃ substrate levels continued to rise, resulting in a ratio of 25(OH)D₃ to 24,25(OH)₂D₃ above 2. The effects of further decreases in Cyp24a1 expression were even more striking in M1-IKO, M1/M21-DIKO, and Cyp27b1-null mice, wherein 24,25(OH)₂D₃ levels decreased rapidly to below the limits of detection (loss of synthesis) while 25(OH)D₃ levels continued to rise almost exponentially (loss of catabolism); the ratio of the two metabolites also continued to increase as well. Finally, as illustrated in the Cyp24a1-null mouse (unable to synthesize 24,25(OH)₂D₃), 25(OH)D₃ reached a peak of over 100 ng/ml, and the ratio of 25(OH)D₃ to 24,25(OH)₂D₃ rose to >50. In contrast to this profile, however, the levels of the terminal degradation product 25(OH)D₃-26,23-lactone decreased in direct proportion to decreasing levels of Cyp24a1 expression. Thus, the ratios of 25(OH)D₃/25(OH)D₃-26,23-lactone in the M1-IKO, M21-IKO, and M1/M21-DIKO mice as well as the Cyp24a1-null and Cyp27b1-null mice all increased to well above 500. The results in the Cyp24a1-null mouse confirm a recent similar measurement in the same mutant mouse strain, a study that also included several humans with inactivating CYP24A1 mutations characteristic of idiopathic infantile hypercalcemia (13, 39, 40). Collectively, these results suggest that an additional role for renal Cyp24a1 may be to modulate actively the circulating levels of 25(OH)D₃ via degradation.

Figure 5.

Schematic representation of Cyp24a1 expression levels correlated to vitamin D₃ metabolites may help predict skeletal health. Shown is the percentage of Cyp24a1 gene expression levels (x axis) plotted against 25(OH)D₃ (gray circles), 24,25(OH)₂D₃ (black squares), and 25(OH)D₃-26,23-lactone (blue triangles) concentrations (y axis) in Cyp24a1-null (C24KO), Cyp27b1-null (C27KO), M1-IKO, M1/M21-DIKO, M21-IKO, and WT littermate control (WT) mice. Skeletal health of the animals is depicted by a green dashed line. Animals to the right of the line are “healthy,” and those to the left are in “poor” skeletal health relative to Ca and P levels as well as circulating concentrations of PTH, FGF23, and 1,25(OH)₂D₃ (6, 29). Error bars, S.E.

Establishing the functional role of the Cyp24-DS2KO regulatory region

Having established a key function of the downstream C24-DS1 regulatory region, we turned to an assessment of the dominant downstream VDR sites located within the C24-DS2 region in C24-DS2KO mice. A peripheral examination revealed a normal size and weight as well as growth pattern in both sexes of these mutant mice relative to their WT littermate counterparts; all of the systemic parameters of these mice that we measured were also equivalent to their WT littermate controls (data not shown). In addition, all mutant mice exhibited a vitamin D metabolite profile including 1,25(OH)₂D₃ levels that were each consistent with that of their WT littermate controls (Table 4). These results led us to predict at best only modest effects on basal expression of Cyp24a1 and perhaps only a modest reduction in response to 1,25(OH)₂D₃ in all tissues. Surprisingly, as seen in Fig. 6A, whereas basal expression of Cyp24a1 was only slightly elevated above normal, this deletion had no effect on either the dose- or time-dependent induction of Cyp24a1 by 1,25(OH)₂D₃ in the kidneys of these mice (Fig. 6B). As expected, however, Cyp24a1 responses to both FGF23 and PTH in the kidney remained fully intact (Fig. 6A). Given the loss of significant VDR-binding sites within the downstream C24-DS2 region, this result suggests that 1,25(OH)₂D₃ action in the kidney is likely mediated exclusively via elements located near the Cyp24a1 gene promoter (C24-PP1). Unexpectedly, however, a similar examination of Cyp24a1 expression and regulation in intestine and bone (two NRTCs tissues) revealed that the loss of the C24-DS2 region significantly reduced the dose- and time-dependent induction of Cyp24a1 by 1,25(OH)₂D₃ (Fig. 6, C and D). The sources of the residual response to 1,25(OH)₂D₃ in these tissues could be derived from either additional VDR binding sites located further downstream of C24-DS2 or through sites located within C24-PP1 or both, although our previous observations using bacterial artificial chromosome clone-transfected bone cells indicate that the Cyp24a1 promoter-proximal-based mechanism is most likely (23). Whereas an effect on basal Cyp24a1 expression in these NRTCs was not seen, an accurate assessment of the gene's basal expression remains difficult, given the low level of Cyp24a1 expression in both of these tissues. Regardless of the source, these results suggest that the deleted downstream region in the C24-DS2KO mouse contributes in a unique manner to the sensitivity of Cyp24a1 to 1,25(OH)₂D₃-mediated up-regulation in these two NRTC tissues. We conclude from this study that the downstream VDR/pCREB regulatory region plays a unique basal but not 1,25(OH)₂D₃-inducible role for Cyp24a1 in the kidney, yet a striking role in mediating the degree of 1,25(OH)₂D₃ response in NRTCs.

Figure 6.

Transcriptional regulation of Cyp24a1 in Cyp24-DS2KO mice. A, gene expression of Cyp24a1 after treatment in 8–9-week-old Cyp24-DS2KO mice (C24-DS2) and their WT littermates (WT) with ethanol/PBS vehicle (Veh, gray, n = 4–6), 10 ng/g bw 1,25(OH)₂D₃ (1,25D₃, black, n = 5–6) for 6 h, 50 ng/g bw FGF23 for 3 h (FGF23, green, n = 5–6), or 230 ng/g bw PTH for 1 h (PTH, blue, n = 5–6) was examined in kidney. B–D, gene expression of Cyp24a1 after treatment in 8–9-week-old Cyp24-DS2KO mice (C24-DS2) and their WT littermates (WT) with ethanol vehicle (Veh, gray, n = 4) or either the indicated amounts of 1,25(OH)₂D₃ (purple, n = 4–5) for 6 h (top row) or 10 ng/g bw 1,25(OH)₂D₃ (purple, n = 4–5) for the indicated periods of time (bottom row) was examined in kidney (B), intestine (C), or bone (D). Data are displayed as relative quantitation (RQ, mean ± S.E. (error bars)) compared with Gapdh. *, p < 0.05, paired t test: C24-DS2 versus WT received the same treatment.

The genomic arrangement of the human CYP24A1 gene in the kidney is similar to that in the mouse kidney

Cyp24a1 is expressed in vitamin D target tissues of all species examined, although basal levels are generally much lower in NRTCs as compared with the kidney. The general regulation by 1,25(OH)₂D₃ in human cells and tissues is also similar, although informative ChIP-Seq scans of the CYP24A1 locus aimed at 1,25(OH)₂D₃ regulation have been limited generally to human NRTC cell lines or primary cells in culture (35, 36, 41–43). Thus, ChIP-Seq analyses of the CYP24A1 locus in human kidney tissue with a focus on regulatory components that might mediate not only 1,25(OH)₂D₃ but PTH and FGF23 action as well have not been conducted. To explore this aspect of CYP24A1 expression relative to our current view of the mouse renal Cyp24a1, we obtained a donated human kidney (26) and conducted both an RNA analysis of CYP24A1 expression and a ChIP-Seq analysis of tissue from the cortex of this organ and compared the genetic and epigenetic landscape with that from the mouse. We assessed residual VDR and pCREB occupancy as well as epigenetic enrichment of H3K4me1 and H3K27ac marks that likely represent signatures of active enhancers. CYP24A1 expression was detected as reported previously (26). The ChIP-Seq tracks in Fig. 7 reveal that both the VDR and pCREB occupy sites near the gene's promoter; the VDR binds downstream of the gene to a subset of sites located distal to the 3′ end of the gene as well. pCREB, on the other hand, while appearing to occupy several unique sites more proximal to the 3′ end of the gene, also co-localizes strongly with virtually all of the more distal sites that bind the VDR. Importantly, each of these novel sites across this downstream locus that bind VDR and pCREB in the human kidney aligns directly with the H3K4me1 and H3K27ac marks that are indicative of the presence of active enhancers. Finally, although linear distances from the CYP24A1 promoter in the human gene differ significantly from those seen in the mouse, the locations and arrangement of these enhancer sites are generally similar. These data in the human kidney together with previous analyses in human cell lines support the idea that the organization of the CYP24A1 gene is highly conserved in both the mouse and human genomes. We entertain the possibility, however, that the actions of PTH, FGF23, and 1,25(OH)₂D₃ within individual enhancers may be consolidated in humans rather than uniquely dispersed across individual enhancers as in the mouse.

Figure 7.

ChIP-Seq analysis of the human kidney reveals conserved genomic occupancy at the CYP24A1 gene locus relative to the mouse. ChIP-Seq analysis of the mouse kidney (top, as in Fig. 2) was contrasted with that derived from isolated human kidney cortex. Overlaid ChIP-Seq data tracks for VDR, pCREB, H3K4me1, and H3K9ac at the Cyp24a1 gene locus from mouse kidney (top) are displayed as either basal or vehicle (yellow, n = 3) or treated for 1 h with PTH or 1,25(OH)₂D₃ (blue, n = 3), as indicated. Overlapping data (vehicle and treatment) appear as green. Bottom, human kidney ChIP-Seq data for pCREB, VDR, H3K4me1, and H3K27ac at the CYP24A1 gene locus are displayed in triplicate (blue, yellow, and pink; overlaps appear as brown). The DS1 and DS2 regions of interest are highlighted in light gray boxes, where the activity of DS2 in NRTCs was established previously (23). Genomic location and scale are indicated (top), and maximum height of tag sequence density for each data track is indicated on the y axis (top left of each track, normalized to input and 10⁷ tags). The direction of transcription is indicated by the arrow, and exons are indicated by boxes.

Discussion

Cyp24a1 is differentially regulated in mouse tissues such that the gene is highly modulated in the kidney by PTH, FGF23, and 1,25(OH)₂D₃, whereas in NRTCs, such as skin, immune cells, bone, and intestine, its regulation is restricted to that conferred largely by 1,25(OH)₂D₃ (1, 7). An important additional feature of Cyp24a1 regulation in the kidney but not in NRTCs is that this gene is controlled by each of the three hormones in a manner directionally opposite of that for Cyp27b1. This reciprocal regulation results in Cyp24a1 suppression by PTH and activation by either FGF23 or 1,25(OH)₂D₃. Although we do not address herein the molecular mechanisms that make this reciprocal regulation of the two genes in the kidney by these hormones possible, these separate properties of Cyp24a1 regulation underscore distinct functional roles for the gene in the kidney versus NRTCs. Accordingly, Cyp24a1 together with Cyp27b1 operate in the former to produce and maintain stable blood levels of active 1,25(OH)₂D₃ and to control adaptive vitamin D metabolism, as illustrated in Fig. 8. In NRTCs, however, Cyp24a1 functions as seen in this figure specifically to modulate the steady-state level of either endocrine or locally derived 1,25(OH)₂D₃ through dynamic control of the hormone's turnover via 24- and 23-hydroxylation, thereby likely playing a determinative role in regulating diverse biological responses.

Figure 8.

Vitamin D metabolism in the kidney and nonrenal tissues. Schematic diagram for the regulation of vitamin D metabolism and serum calcium and phosphate homeostasis in the kidney (top) and nonrenal target cells (NRTCs, bottom). Our genetic models (black) and previously existing models (gray) are overlaid on or near the pathways they disrupt. This figure was derived from our earlier publication (7).

Cyp24a1 participates in the metabolism of vitamin D through its actions to control the level of circulating 1,25(OH)₂D₃ and to regulate its intracellular level in vitamin D target cells as well. Thus, Cyp24a1 is expressed in the kidney and almost ubiquitously in numerous additional tissues in vivo. These features underscored the need for an in vivo study of the regulation of Cyp24a1, wherein multiple target tissues and genes could be examined at a genomic level in the context of a potentially complex systemic, metabolic, and/or skeletal phenotype that we suspected might emerge as a result of altered mineral regulation. Thus, our approach involved the identification of potential regulatory regions in the Cyp24a1 gene locus using ChIP-Seq analysis followed by CRISPR/Cas9-mediated enhancer deletion together with retention-of-function (ROF)/loss-of-function (LOF) studies in the mouse. There are multiple advantages to this in vivo approach, not the least of which is the ability to examine gene regulation in response to hormonal challenge in multiple tissues. Several disadvantages are also apparent, however, including the fact that the preparation of LOF/ROF deletions in mice is time-consuming and additional resolution is certain to require additional rounds of genomic deletion. A second disadvantage is that both the biological and regulatory phenotypes that emerge are likely to be highly complex, although this complexity can provide novel insight into the gene's regulation and function, as identified herein.

Our initial studies revealed a number of unexpected features of Cyp24a1 regulation. Thus, as summarized above, we found that in concert with the well-known enhancer activity located near the Cyp24a1 gene promoter (C24-PP1), a complex set of enhancers was also located downstream of the gene. Importantly, the more proximal subset (C24-DS1) displayed dispersed residual pCREB occupied sites, was present only in the kidney, and mediated unique regulatory properties relative to PTH and FGF23. The more distal subset (C24-DS2), on the other hand, was comprised of several clusters of VDR-binding sites that highlighted an interesting redundancy and was present in both the kidney and in NRTCs, yet appeared largely active only in nonrenal tissue types, where it increased sensitivity to the 1,25(OH)₂D₃ hormone. These properties define a fundamental underlying genomic mechanism through which Cyp24a1 regulation by PTH and FGF23 is restricted to the kidney, whereas the gene's regulation by 1,25(OH)₂D₃ is mediated through differential utilization of regulatory elements present in both the kidney and in NRTCs. Interestingly, the more distal downstream VDR interactive C24-DS2 segment was identified originally by ChIP-Seq analysis in several nonrenal mouse and human cell lines (23). In these earlier studies, we also showed using mouse and human bacterial artificial chromosome clones and stable cell transfections that both regulatory regions contributed functionally to up-regulation by 1,25(OH)₂D₃ in vitro, a finding that has now been duplicated in these in vivo studies. Our current findings will now enable us to focus on additional dissection to resolve the downstream sites of PTH versus FGF23 action, to confirm that pCREB is the mediator of the former hormone, and to identify the transcription factor(s) and perhaps the signaling pathway(s) that is involved in FGF23 activity at Cyp24a1.

The observation that Cyp24a1 is reciprocally regulated by PTH and FGF23 in the kidney is equally important because this feature provides the distinctive homeostatic linkage between Cyp24a1 and Cyp27b1 that governs the regulation of circulating endocrine 1,25(OH)₂D₃. Our previous studies and this paper show that whereas PTH induces Cyp27b1, it suppresses Cyp24a1. Similarly, whereas both FGF23 and 1,25(OH)₂D₃ suppress Cyp27b1, they induce Cyp24a1. This reciprocal regulation results in a mechanism through which 1,25(OH)₂D₃ can be maintained at physiological levels. Interestingly, this reciprocal regulation of Cyp27b1 and Cyp24a1 in the kidney is also operable for preserving blood 1,25(OH)₂D₃ levels when either genetic, epigenetic, or pathophysiological processes cause a sustained disturbance in either Cyp27b1 or Cyp24a1 expression in the kidney. Thus, when the production of 1,25(OH)₂D₃ is insufficient due to a fixed down-regulation of renal Cyp27b1 expression, homeostatic mechanisms involving PTH and FGF23 decrease renal Cyp24a1 expression, thereby reducing 1,25(OH)₂D₃ turnover, which results in the preservation of circulating hormone. This leads to higher-than-expected levels of 1,25(OH)₂D₃, prompting the hormone's characterization in patients as “inappropriately high” relative to clinical expectations. Indeed, pathophysiological reductions in Cyp27b1 are known to arise as a result of genetic abnormalities that reduce or inhibit expression of Cyp27b1 (alteration in expression due to SNPs, noncoding genomic deletions, or gene duplication) or reduce 1,25(OH)₂D₃ levels due to mutations within the Cyp27b1 gene itself (vitamin D resistance, type 1) (44–46). Each of these situations has been identified in humans and replicated in genetic mouse models as well (47). The recent generation of the M1-IKO and the M1/M21-DIKO mouse strains has revealed similar findings (7, 26). In the current studies, we show that this co-regulation is also homeostatically linked in reverse when Cyp24a1 expression is reduced or lost, as exemplified in the C24-DS1KO mouse. Accordingly, the suppression of Cyp24a1 expression due to the strong reduction in basal activity leads to a coordinated down-regulation of Cyp27b1 that avoids a progressive elevation in blood 1,25(OH)₂D₃ levels that occurs as a result of reduced turnover. Importantly, the dynamic changes in the levels of PTH and FGF23 that affect vitamin D metabolism and thus mineral homeostasis have little impact on the local expression of Cyp24a1 or Cyp27b1 in NRTCs and are therefore less likely to alter the biological activity of 1,25(OH)₂D₃ in those cell types.

The concept of differential regulation also raises several questions as to why the control of Cyp24a1 expression by 1,25(OH)₂D₃ differs in the kidney versus NRTCs, where the former is modulated exclusively through the promoter-proximal region (C24-PP1), whereas the latter is regulated by both this region and the downstream distal segment (C24-DS2). The kidney represents the major endocrine source of secreted 1,25(OH)₂D₃, so it is possible that up-regulation of Cyp24a1 must be carefully insulated from this local production that could conflict with that of downstream PTH and/or FGF23 regulation. It is also possible that additional regulatory controls unique to the Cyp24a1 promoter region could oppose local renal 1,25(OH)₂D₃ action and thus blunt the up-regulatory activity of the hormone as well. This could occur because renal Cyp24a1 is expressed largely in cells of the distal convoluted tubule, whereas Cyp27b1 synthesis is predominantly a proximal tubule function (1). An additional question is raised as to how Cyp24a1 promoter-proximal selectivity (at C24-PP1) can be exerted by 1,25(OH)₂D₃ in the kidney, when VDR binding is observed in this tissue across both this region and within C24-DS2, whereas hormone activity is limited to the former. Is it possible that this downstream region contributes to renal Cyp24a1 expression under physiological circumstances different from those tested herein? 1,25(OH)₂D₃, on the other hand, is a strong regulator of Cyp24a1 in NRTCs through both promoter-proximal C24-PP1 and downstream C24-DS2 regions (26). This increased sensitivity to circulating levels of 1,25(OH)₂D₃ may underlie a dynamic mechanism for controlling the set point for intracellular 1,25(OH)₂D₃ activity in different cell types. Perhaps more importantly, this elevated sensitivity to 1,25(OH)₂D₃ may also provide for increased cellular resistance to potentially toxic circulating levels of renal 1,25(OH)₂D₃ if such circumstances arise. Although these issues are all speculative, they do provide a framework for further investigation into the possible functions of Cyp24a1 in kidney and in different NRTCs.

An investigation of the relationship between the blood concentrations of 25(OH)D₃, 24,25(OH)₂D₃, and 25(OH)D₃-26,23-lactone and the level of renal Cyp24a1 expression in a series of mutant mouse strains revealed that as Cyp24a1 levels modestly decreased, both 24,25(OH)₂D₃ and 25(OH)D₃ levels tended to rise such that the 25(OH)D₃/24,25(OH)₂D₃ ratio was maintained between 1.6 and 2. As Cyp24a1 expression was reduced further, however, the levels of 24,25(OH)₂D₃ and 25(OH)D₃ diverged, resulting in an increasing ratio that reached >100. In contrast, however, a direct proportionality between the concentration of 25(OH)D₃-26,23-lactone and the entire range of Cyp24a1 expression was maintained, leading to ratios between 25(OH)D₃ and 25(OH)D₃-26,23-lactone at low levels of Cyp24a1 expression that were much higher than 500. Although it seems counterintuitive that 24,25(OH)₂D₃ levels should rise with declines in Cyp24a1 expression, this differing relationship between Cyp24a1 levels and 24,25(OH)₂D₃ versus 25(OH)D₃-26,23-lactone occurs because 24,25(OH)₂D₃ represents the initial step in the catabolism of 25(OH)D₃ via C24-hydroxylation, whereas the 25(OH)D₃-26,23-lactone represents the terminal step in the catabolism of 25(OH)D₃ via C23-hydroxylation. This was determined recently in a study where WT mice were given high doses of 24,25(OH)₂D₃ and metabolism could be traced through all C24-hydroxylation pathway metabolites to calcioic acid, the unique final product of this pathway (48). Following Cyp24a1 ablation, however, whereas downstream metabolites were undetectable, 24,25(OH)₂D₃ accumulated 4-fold, highlighting the important role of CYP24A1 in the further catabolism of both 24,25(OH)₂D₃ as well as 25(OH)D₃. Thus, our correlative secondary analysis suggests that in the face of lowered Cyp24a1 expression, increased 24,25(OH)₂D₃ levels are likely due to an accumulation of uncatabolized 24,25(OH)₂D₃, which is still being formed from 25(OH)D₃, albeit at a slower rate than seen in WT mice. Accordingly, the peak levels of 24,25(OH)₂D₃ that we observed in the M21-IKO mouse as well as in the current C24-DS1KO mouse likely represent the tipping point rates of 24,25(OH)₂D₃ production relative to further degradation by CYP24A1. The direct proportionality between Cyp24a1 expression and terminal 25(OH)D₃-26,23-lactone levels, however, allowed several of us to conclude that the rise in the ratio of 25(OH)D₃ to 25(OH)D₃-26,23-lactone is a much more sensitive determinant of Cyp24a1 expression than that of 25(OH)D₃ to 24,25(OH)₂D₃. Equally important in this study, whereas Cyp24a1 was strikingly down-regulated in the kidney of each of these mouse strains, the expression of Cyp24a1 remained unaffected in NRTCs, indicating that the enzyme in these tissues was likely not involved.

The overall profile of data obtained from these multiple strains of mice also revealed an important additional concept with regard to renal Cyp24a1 expression. Clearly, as can be seen in the graph in Fig. 5, the maintenance of appropriate physiological levels of 25(OH)D₃ is fully dependent upon the normal expression and activity of renal Cyp24a1. Thus, as Cyp24a1 expression becomes progressively suppressed and 24,25(OH)₂D₃ production is eventually lost, 25(OH)D₃ rises dramatically, leading to an increase in the ratio of the two metabolites. Because the status of each of these mice is health-adverse, characterized by hypocalcemia, hypophosphatemia, hyperparathyroidism, low levels of both FGF23 and 1,25(OH)₂D₃, and skeletal deformities, elevated ratios appear indicative of significant pathological status due to decreased Cyp24a1 expression and/or activity (7, 26). Thus, from a basic perspective, Fig. 5 also reveals an additional important fundamental concept that renal Cyp24a1 modulates not only 1,25(OH)₂D₃ but 25(OH)D₃ levels as well. Accordingly, whereas it has been widely recognized that the hepatic synthesis of 25(OH)D₃ via the activities of several different enzymes represents an unregulated step in the initial activation of vitamin D₃, it seems clear that the concentration of 25(OH)D₃ in the blood is indeed actively regulated via turnover through the catabolic activity of renal Cyp24a1. The physiologic rationale underlying this regulation to limit 25(OH)D₃ production levels is unclear, although it is worth noting that both the level and the potential transcriptional activity of 24,25(OH)₂D₃ and its further metabolic degradation products is much lower than that for 25(OH)₂D₃. Accordingly, the conversion of 25(OH)D₃ may represent a mechanism through which the potential toxicity of 25(OH)D₃ may be avoided when uncontrolled hepatic production is increased, by analogy with the catabolism of 1,25(OH)₂D₃ to slightly less active 1,24,25(OH)₃D₃. Regardless of this, it is apparent that in addition to its contribution in controlling 1,25(OH)₂D₃ levels in the blood, renal Cyp24a1 plays an active role in limiting and therefore maintaining physiologically appropriate levels of the 25(OH)D₃ substrate as well.

Finally, previous studies of the Cyp24a1 locus revealed the presence of downstream VDR clusters active in both mouse and human NRTC cell lines (23) but not in the kidney. We therefore assessed in final studies the presence of such a control mechanism in the human kidney and identified both similarities and differences. We note that whereas the downstream segments of C24-DS1 and C24-DS2 are both present in the human kidney, based upon both VDR and pCREB binding, the locations of these sites and the degree of overlap between the two appear to be much greater. Whether and how these two factors might influence each other's activity is unclear, although the actions of both PTH and FGF23 on Cyp24a1 expression are directionally opposite each other. In any event, these observations of the human kidney suggest an increase in the complexity of the arrangement relative to the mouse and highlight the necessity in the future for examining the individual roles of each enhancer in the regulation of human CYP24A1 expression.

In summary, we have identified a chromatin-based mechanism that mediates the selective regulation of Cyp24a1 in the kidney and in NRTCs by PTH, FGF23, and 1,25(OH)₂D₃. This differential regulation highlights the important contributory role that Cyp24a1 plays in sustaining blood levels of endocrine 1,25(OH)₂D₃ and reveals how the coordinated reciprocal regulation of both Cyp27b1 and Cyp24a1 contributes to this process. It also reveals the distinctive dual roles of Cyp24a1 in the kidney to regulate the circulating levels of both 1,25(OH)₂D₃ and 25(OH)D₃ and in NRTCs to modulate 1,25(OH)₂D₃ turnover and thus cellular responses in a manner that is largely independent of the hormones that control mineral metabolism. Further studies will be necessary to define the numerous individual sites of action of PTH and FGF23 in the kidneys and to understand the mechanisms that control the activation of the transcription factors involved and indeed in the case of FGF23 the identity of the factor(s) itself.

Experimental procedures

Reagents

The following reagents were used for in vivo injections. 1α,25(OH)₂D₃ was obtained from SAFC Global (Madison, WI), PTH(1–84) (human) was obtained from Bachem (H-1370.0100 Torrence, CA), and mouse FGF23 was from R&D Systems (FGF23, 2629-FG-025, Minneapolis, MN). Antibodies used for ChIP-Seq analysis of VDR (C-20, sc-1008, lot H1216) were purchased from Santa Cruz Biotechnology, Inc. H3K4me1 (ab8895, lot GR283603-1), H3K27ac (ab4729, lot GR3187597-1), and H3K36me3 (ab9050, lot GR273247-1) were purchased from Abcam (Cambridge, MA). pCREB (Ser-133) (06-519, lot 2762242) and H3K9ac (06-942, lot 2664263) were purchased from Millipore Corp. (Billerica, MA). Traditional genotyping PCR was completed with GoTaq (Promega, Madison, WI), and all real-time quantitative PCR (qPCR) was completed with the StepOnePlus using TaqMan for gene expression assays (Applied Biosystems, Foster City, CA). Primers were obtained from IDT (Coralville, IA).

Gene expression

Dissected tissues were frozen immediately in liquid nitrogen and stored at −80 °C. Frozen tissues were homogenized in TRIzol reagent (Life Technologies, Inc.), and RNA was isolated as per the manufacturer's instructions. 1 μg of isolated total RNA was DNase-treated, reverse-transcribed using the High Capacity cDNA Kit (Applied Biosystems), and then diluted to 100 μl with RNase/DNase-free water. qPCR was performed using primers specific to a select set of differentially expressed genes by TaqMan analyses. TaqMan Gene Expression probes (Applied Biosystems) were used for RT-PCR and are found in Table 5.

Table 5.

TaqMan assays, guide RNAs, and genotyping primers

Name	Sequence/Source	Reference
TaqMan primers
Gapdh	Applied Biosystems	4352339E
Cyp27b1	Applied Biosystems	Mm01165918
Cyp24a1	Applied Biosystems	Mm00487244
Vdr	Applied Biosystems	Mm00437297
Primers for CRISPR
Cyp24-DS Guide 1	TAGGGACAAGGCATCAGACG-AGG	This work
Cyp24-DS Guide 2	AGAGTCGAGCGGAAATGTGC-AGG	This work
Cyp24-DS Guide 3	GTTTATAGAATCCAGCTTGG-AGG	This work
Primers for genotyping
Cyp24-DS1 span F	GGATGAACCAAGAAGGTGCTGTG	This work
Cyp24-DS1 span R	CTGGCGTAGTGAGGAAAACCT	This work
Cyp24-DS1 internal F	GATTTTGTTAGACGTGGCAGGG	This work
Cyp24-DS1 internal R	CTTCCTTTGGACAAGCATCGTG	This work
Cyp24-DS2 span F	ATTGCCTAACCAGCCCTCTT	This work
Cyp24-DS2 span R	TGTGGAGTGCTGGAATTTGC	This work
Cyp24-DS2 internal F	TTTAGAAGTGAGGGCTGCCA	This work
Cyp24-DS2 internal R	TTTGCTCCTGTTGACACTGC	This work

ChIP-Seq

ChIP was performed using antibodies listed under “Reagents.” ChIP was performed as described previously with several modifications (7, 49). Human kidney samples were obtained from the University of Wisconsin-Madison Organ Procurement Organization with approval as recently published (26). The isolated DNA (or input DNA acquired prior to precipitation) was then validated by qPCR and further prepared for ChIP-Seq analysis. ChIP-Seq libraries were prepared as described previously (7, 50). All human data were mapped to the hg19 genome build, and mouse data were mapped to the mm9 build. All human and mouse kidney ChIP-Seq data have been deposited in the Gene Expression Omnibus (GSE129585 and GSE133025) (26).

CRISPR-generated and transgenic mice

The guides used for CRISPR/Cas9-mediated genome editing were optimized for the least number of potential off-target sites and fewest sites within coding exons using the Zhang laboratory CRISPR Design tool and cross-referenced with the Liu laboratory CRISPR-DO. The guides (see Table 5) were annealed and cloned into plasmids pX330 or pX458 obtained from the Zhang laboratory via Addgene (Cambridge, MA) as described recently (51). Resulting PCR products were then transcribed in vitro utilizing the T7 MEGAshortscript kit (Life Technologies, Inc.) (52). The mixture of 50 ng/μl of the produced RNA guides and 40 ng/μl of Cas9 protein in injection buffer (5 mm Trizma base, 5 mm Tris-HCl, 0.1 mm EDTA, pH 7.4) was injected into the pronucleus of 1-day fertilized embryos isolated from hyperovulating female C57BL/6 mice as described previously (53). Embryos were then implanted into recipient females by the University of Wisconsin-Madison Biotechnology Genome Editing and Animal Models Core. The resultant pups were genotyped with spanning primers (Table 5), cloned, and sequenced. The top 10 predicted potential off-target sites were examined by PCR and sequencing analysis. Founder mice for DS1 (3 HETs) and DS2 (2 HETs) were retained for outbreeding and experimentation.

Animal studies

Genetically modified mice were outbred with C57BL/6 mice (Jackson Laboratory, Bar Harbor, ME) as heterozygotes. Mice were housed in high-density ventilated caging in the Animal Research Facility of the University of Wisconsin-Madison under 12-h light/dark cycles at 72° F and 45% humidity. All mice used in this study were maintained on a standard rodent chow diet (5008, Lab Diet, St. Louis, MO), aged 8–9 weeks, and then outcrossed with WT C57BL/6J mice 5 or more generations unless otherwise indicated. All experiments and tissue collections were performed in the procedure rooms in the Research Animal Facility of the University of Wisconsin-Madison. All animal studies were reviewed and approved by the Research Animal Care and Use Committee of the University of Wisconsin-Madison under Protocol A005478. Homozygous mice and littermate controls were generated by HET × HET cross and evaluated for their systemic as well as skeletal phenotypes and for both basal and hormone-regulated expression of Cyp24a1 in the kidney and in NRTC tissues. Animals were subjected to intraperitoneal injection of 10 ng/g body weight (bw) 1,25(OH)₂D₃ (in propylene glycol), 230 ng/g bw PTH (1–84) (in PBS), 50 ng/g bw FGF23 (in PBS + 0.1% BSA), or vehicle (EtOH or PBS). Animals were sacrificed, and tissues were collected 1 h after PTH injection, 3 h after FGF23 injection, and 6 h after 1,25(OH)₂D₃ injection. Unless otherwise indicated, all experiments were conducted with equal numbers of males and females (n ≥ 6). There was no difference in male and female animals unless otherwise indicated. Original data documenting the expression of Cyp24a1 from the kidneys of M1-IKO, M21-IKO, M1/M21-DIKO, Cyp27b1-null, and Cyp24a1-null mouse strains and their individual littermate controls as well as summaries of individual phenotypes and vitamin D metabolite profiles were derived from our recently published papers (7, 26). Expression of additional genes that included Cyp27b1 and Vdr was also assessed. The Cyp24a1-null mouse strain was a kind gift from Professor Rene St-Arnaud (Shriner's Hospital/McGill University, Montreal, Canada).

Blood chemistry

Cardiac blood was collected at the time of sacrifice. Collected blood was split into serum- or EDTA-treated plasma and incubated at room temperature for 30 min, followed by centrifugation at 6,000 rpm for 12 min (twice) to obtain serum or EDTA plasma. Serum calcium and phosphate levels were measured using the QuantiChrom^TM calcium assay kit (catalog no. DICA-500, BioAssay Systems, Hayward, CA) and the QuantiChrom^TM phosphate assay kit (catalog no. DIPI-500, BioAssay Systems). Circulating intact FGF23 and PTH were measured in EDTA plasma via a mouse/rat FGF23 (Intact) ELISA kit (catalog no. 60-6800, Immutopics, San Clemente, CA) and a Mouse PTH(1–84) ELISA kit (catalog no. 60-2305, Immutopics), respectively.

Quantification of serum vitamin D metabolites

Serum 25(OH)D₃, 24,25(OH)₂D₃, 1,25(OH)₂D₃, and 1,24,25(OH)₃D₃ were quantified by LC-MS/MS using previously published methods (15, 39, 40), as recently reported (7).

BMD

At 8–9 weeks of age, BMDs of the CRISPR-generated mice and their WT littermates were measured and analyzed by dual X-ray absorptiometry with a PIXImus densitometer (GE-Lunar Corp., Madison, WI) as described previously (54).

Statistical evaluation

Data were analyzed using GraphPad Prism 8 software (GraphPad Software, Inc., La Jolla, CA) and in consultation with the University of Wisconsin Statistics Department. All values are reported as mean ± S.E., and differences between group means were evaluated using one-way ANOVA, two-way ANOVA, or Student's t test as indicated in the figure legends.

Author contributions

J. W. P. and M. B. M. conceptualization; M. B. M., S. M. L., A. H. C., N. A. B., and M. K. investigation; J. W. P. and M. B. M. writing-original draft; N. A. B., S. M. L., M. K., and G. J. writing-review and editing; M.B.M. data curation.