Memo1 gene expression in kidney and bone is unaffected by dietary mineral load and calciotropic hormones.

Mediator of cell motility 1 (MEMO1) is a ubiquitously expressed modulator of cellular responses to growth factors including FGF23 signaling, and Memo1-deficient mice share some phenotypic traits with Fgf23- or Klotho-deficient mouse models. Here, we tested whether Memo1 gene expression is regulated by calciotropic hormones or by changing the dietary mineral load. MLO-Y4 osteocyte-like cells were cultured and treated with 1,25(OH)2 -vitamin D3 . Wild-type C57BL/6N mice underwent treatments with 1,25(OH)2 -vitamin D3 , parathyroid hormone, 17β-estradiol or vehicle. Other cohorts of C57BL/6N mice were fed diets varying in calcium or phosphate content. Expression of Memo1 and control genes was assessed by qPCR. 1,25(OH)2 -vitamin D3 caused an acute decrease in Memo1 transcript levels in vitro, but not in vivo. None of the hormones tested had an influence on Memo1 transcripts, whereas the assessed control genes reacted the expected way. Dietary interventions with calcium and phosphate did not affect Memo1 transcripts but altered the chosen control genes' expression. We observed that Memo1 was not regulated by calciotropic hormones or change in mineral load, suggesting major differences between the regulation and physiological roles of Klotho, Fgf23, and Memo1.

INTRODUCTION

Memo1 is an evolutionary conserved protein in all kingdoms of life that has shown intracellular expression in cytoplasma and nucleus (Haenzi et al., 2014; Moor, Haenzi, et al., 2018; Schlatter et al., 2012). A conditional and inducible knockout mouse model with postnatal deletion of exon 2 of the Memo1 gene has resulted in a syndrome of aging and premature death with traits such as elevated calcemia, elevated FGF23 and 1,25(OH)2‐vitamin D3, bone disease, lung emphysema, atrophy of subcutaneous fat, insulin hypersensitivity, and renal insufficiency (Haenzi et al., 2014; Moor, Ramakrishnan, et al., 2018). This phenotype significantly overlaps with phenotypes of mouse models deficient in KLOTHO (Kuro‐o et al., 1997) or FGF23 (Shimada et al., 2004), two proteins which have tremendously reshaped our understanding of the regulation of calcium and phosphate metabolism by the kidney and bone and to a lesser extent also the intestine (Hu, Shiizaki, Kuro‐O, & Moe, 2013; Moor & Bonny, 2016).

In addition, evidence from cell culture experiments investigating co‐localization and phosphorylation status of adaptor proteins suggested that mediator of cell motility 1 (MEMO1) protein participates in and modulates a signaling cascade involving FGF23 and the FGFR (Haenzi et al., 2014).

Serum analyses of Klotho and Fgf23‐deficient mice showed excessive 1,25(OH)2‐vitamin D3 levels, a finding that has been variably found in Memo1‐deficient mice depending on the genetic background (Haenzi et al., 2014; Moor, Ramakrishnan, et al., 2018). The promoters of Fgf23 and Klotho both contain vitamin D response elements (VDRE) (Forster et al., 2011; Orfanidou, Malizos, Varitimidis, & Tsezou, 2012). FGF23 secretion is increased by parathyroid hormone (PTH) (Lavi‐Moshayoff, Wasserman, Meir, Silver, & Naveh‐Many, 2010) and 17β‐estradiol (Carrillo‐Lopez et al., 2009). Regarding the regulation of Memo1, a transcriptomic analysis of rat pineal glands has detected increased Memo1 transcripts upon synthetic estrogen treatment compared to controls (Deffenbacher & Shull, 2006), highlighting a potential endocrine regulation of the transcription of the Memo1 gene. Here, we, therefore, tested the hypothesis that Memo1 expression can be regulated by minerals or calcitropic stimuli.

METHODS

Cell culture

The mouse long bone osteocyte‐Y4 cell line (MLO‐Y4) was kindly provided by Lynda Bonewald (Kato, Windle, Koop, Mundy, & Bonewald, 1997). MLO‐Y4 cells were maintained in culture in minimal essential medium alpha‐modified, alpha‐MEM (Gibco by Life Technologies), containing 2.5% heat‐inactivated calf serum (Sigma), 2.5% heat‐inactivated fetal bovine serum (Sigma), and 1% penicillin/streptomycin (Invitrogen by Life Technologies). Serum heat inactivation was carried out in water bath at 56°C for 30 min. Cells were cultured on rat‐tail type I collagen (Invitrogen by Life Technologies). For vitamin D stimulation, cells were kept in serum‐free medium supplemented with 10 nM 1,25(OH)2‐vitamin D3 (Sigma) or ethanol vehicle for 24 hr.

Animal experiments

C57BL/6N mice were obtained from Janvier. Mice were held in a conventional animal facility with up to six animals per cage and they were fed a standard laboratory chow (Kliba Nafnag TS3242; calcium 1%, phosphorus 0.65%, magnesium 0.23%, vitamin D 1,600 IU/kg, vitamin A 27,000 IU/kg, vitamin E 150 mg/kg, protein 18.8%, crude fat 5.6%, crude fiber 3.5%, lysine 1.1%; KLIBA) unless specified otherwise and were kept on 12/12 (experimental) or 14/10 (breeding) light–dark cycles. All animal experimental protocols were approved by the State Veterinary Service of the Canton de Vaud, Switzerland. For all mouse studies, sample sizes were considered based on previous results in our laboratory.

Dietary interventions

For studies of mineral metabolism, specifically designed diets and hormones were used. Calcium diet challenge experiments were carried with five male C57BL/6N mice per condition in their home cage, all aged 12 weeks. Mice were randomly allocated to be fed modifications of KLIBA 2222 diet containing either 0.17% (low calcium diet), 0.82% (normal calcium diet) or 1.69% (high calcium diet) (KLIBA, 2222) over 7 days. All calcium diets contained phosphorus 0.8%, vitamin A 4,000 IU/kg, vitamin D 1,000 IU/kg, vitamin E 100 mg/kg, protein 18%, and crude fat 7%, lysine 14 g/kg (Kliba, 2222). For a phosphate diet challenge, three groups of five male C57BL/6N mice all aged 13–14 weeks were kept in metabolic cages and randomly allocated to be fed diets containing low (0.2%), intermediate (0.8%), or high (1.5%) phosphate content over 7 days. The phosphate diets were modifications of KLIBA 2222 diet (calcium 1.2%, vitamin A 4,000 IU/kg, vitamin D 1,000 IU/kg, vitamin E 100 mg/kg, protein 18%, crude fat 7%, and lysine 14 g/kg).

Hormone injections

All hormone injections were performed in the home cage of the mice after random treatment allocation of individual ear‐marked mice. For 1,25(OH)2‐vitamin D3 treatment, male C57BL/6N mice aged 13–15 weeks were subcutaneously injected with 2 μg/kg body weight 1,25(OH)2‐vitamin D3 (Sigma D1530) in ethanol 1% in NaCl 0.9%. The dose and application route was the same as previously used in our laboratory. Control mice were injected with 1% (v/v) ethanol in NaCl 0.9%. Mice were sacrificed 6 hr after injection.

For PTH treatment, male C57BL/6N mice aged 12–13 weeks were subcutaneously injected with 80 μg/kg body weight human PTH fragment 1–34 (hPTH1‐34) (Sigma P3796) in NaCl 0.9% or NaCl 0.9% alone as vehicle and sacrificed 2 hr after injection. The dose and application route used was determined from the literature (Kramer, Loots, Studer, Keller, & Kneissel, 2010).

For estradiol treatment, male C57BL/6N mice aged 16 weeks received one daily subcutaneous injection of 15 μg 17β‐Estradiol (Sigma E8875) in ethanol 0.1% (v/v) in NaCl 0.9% for five consecutive days and were sacrificed 4 hr after the last injection. The dose per body weight and the drug application route was derived from the literature, as shown Van Abel et al. (2002) to induce Trpv5 expression. Control mice were subcutaneously injected with 0.1% ethanol in NaCl 0.9% for 5 days.

Mouse dissection

For euthanasia, mice were intraperitoneally injected with 0.1 mg/gBW of ketamine (Ketanarkon 100 Vet., Streuli) and 0.02 mg/gBW of xylazine (Rompun, Bayer), followed by terminal exsanguination by orbital puncture under full anesthesia and/or by cervical dislocation. Organs were collected, kidneys were cut in half, and organs were snap frozen in liquid nitrogen immediately, followed by storage at −80°C until further use.

RNA extraction

RNA was extracted using TRI reagent (Applied Biosystems by Life Technologies) according to manufacturer's instructions. RNA pellets were dried and dissolved in RNase‐free H2O. RNA concentration was measured photometrically using NanoDrop (NanoDrop 2000, Thermo Fisher Scientific). RNA A260/A280 ratio was assessed and each RNA sample was visualized on a 1% agarose gel.

RNA was reverse transcribed to cDNA using the PrimeScript RT reaction kit (Takara Bio Inc). RNA input quantities per sample were 1–2 μg for bone, 500 ng for kidney or 1 μg of MLO‐Y4 RNA. The resulting cDNA mix was diluted 2–12x depending on tissue type.

qPCR

For quantitative gene transcript expression analysis, 2 μL of cDNA was used for SYBR Green qPCR (Applied Biosystems by Life Technologies) on a 7500 Fast machine (Applied Biosystems). Samples were run in triplicate in 20 μL total volume for each gene, and actin or GAPDH was used for normalization. Melting curves were obtained for every run. Program settings were: 95°C during 20 s, 40 cycles (95°C 3 s, 60°C 30 s), and for melting curve stage: 95°C 15 s, 60°C 1 min, rising at 1% ramp speed to 95°C (15 s), and 60°C 15 s. Data were analyzed using the delta‐delta CT method. Primers were ordered from Microsynth (Switzerland) and sequences are shown in Table 1. All amplified products were visualized on agarose gels.

Table 1

Primers used for qPCR

Oligonucleotide	5'‐sequence‐3'
Memo1 forward	GCTGCCCATGCTTACAAACAA
Memo1 reverse	AGAGTGCACATCGAGACAGG
Rankl forward	GTCTGTAGGTACGCTTCCCG
Rankl reverse	CATTTGCACACCTCACCATCAAT
Bglap forward	CCGCCTACAAACGCATCTATG
Bglap reverse	GCTGCTGTGACATCCATACTTG
Phex forward	GTGCATCTACCAACCAGATACG
Phex reverse	TCTGTTCCCCAAAAGAAAGG
Slc34a3 forward	CCTACCCCCTCTTCTTGGGT
Slc34a3 reverse	AGAGCAACCTGAACTGCGAA
F3 forward	ACCTGGGCCTATGAAGCAAA
F3 reverse	GTTGGTCTCCGTCTCCATGAA
Cyp27b1 forward	ATGTTTGCCTTTGCCCAGA
Cyp27b1 reverse	GACGGCATATCCTCCTCAGG,
Cyp24a1 forward	GAAGATGTGAGGAATATGCCCTATTT
Cyp24a1 reverse	CCGAGTTGTGAATGGCACACT
beta‐actin forward	GTCCACCTTCCAGCAGATGT
beta‐actin reverse	AGTCCGCCTAGAAGCACTTGC

Data analysis

Human Memo1 promoter sequences were analyzed in silico using the UCSC Genome Browser and Serial Cloner 2.6.1.

Data from experiments with two independent groups were analyzed by t test or Mann–Whitney U test. For comparison of three groups, Kruskal–Wallis test was used with Bonferroni's Multiple Comparison posttest. All statistical analyses were conducted using GraphPad PRISM 5.03. Two‐sided p < .05 were considered significant.

RESULTS

Memo1‐deficient mice resemble by some traits Klotho mutant and FGF23 KO mice (Haenzi et al., 2014), and the promoters of Klotho and FGF23 harbor regulatory sequences that can be bound by vitamin D receptors (Forster et al., 2011; Orfanidou et al., 2012). For these reasons, we determined the regulation of Memo1 gene expression by minerals and calciotropic hormones. We have previously shown that MLO‐Y4 osteocyte‐like cell line expressed Memo1 transcripts and protein (Moor, Ramakrishnan, et al., 2018).

Memo1 is diminished by 1,25(OH)2‐vitamin D3 in vitro but not in vivo

An in silico promoter analysis of the human Memo1 gene revealed a conserved CpG island (Gardiner‐Garden & Frommer, 1987) sequence in the 1,000 bases in 5’ direction of transcription start site of Memo1 that was considered as the putative promoter sequence (Figure S1). In a screen for published VDREs, we identified two incidences of a negative VDRE with the sequence 5′‐GCTTTCC‐3′ (Towers, Staeva, & Freedman, 1999). Five VDRE sequences reported elsewhere (Calle, Maestro, & Garcia‐Arencibia, 2008; McGaffin & Chrysogelos, 2005; Roff & Wilson, 2008) were undetectable on either sense or antisense strand. Therefore, we proceeded to experimentally investigate the effects of stimulation with 10 nM 1,25(OH)2‐vitamin D3 on gene expression in MLO‐Y4 cells (Figure 1). Known vitamin D‐dependent transcripts were first assessed. Transcripts of Cyp24a1 encoding a vitamin D inactivating hydroxlyase (Figure 1a) and of osteoclast regulator Rankl (Figure 1c) were increased upon 1,25(OH)2‐vitamin D3 treatment, whereas transcripts of FGF23 regulator Phex and expression of bone‐derived hormone osteocalcin/bone gamma‐carboxyglutamate (Bglap) were diminished (Figure 1b,d). Memo1 transcripts were reduced by 20% upon 1,25(OH)2‐vitamin D3 treatment (Figure 1e).

Figure 1

Transcriptional responses to 1,25(OH)2‐vitamin D3 in MLO‐Y4 osteocyte‐like cells. 1,25(OH)2‐vitamin D3 treatment significantly increased Cyp24a1 transcripts (a), decreased Phex transcripts (b), increased Rankl transcripts (c), decreased Bglap (osteocalcin) transcripts (d), and decreased Memo1 (e) transcripts in MLO‐Y4 cells. n = 6 per condition, *p < .05 (t test)

Next, we intraperitoneally injected 1,25(OH)2‐vitamin D3 in mice. Six hours post‐injection we harvested the kidney and tibia of these animals and investigated the mRNA levels of Memo1. Memo1 RNA levels remained unchanged in the kidney (Figure 2a) and tibia (Figure 2b) compared to control animals injected with vehicle only. As experimental controls, we chose Cyp24a1 and Fgf23. Renal transcripts of Cyp24a1, the gene encoding the vitamin D inactivating enzyme cytochrome P450 24a1, were upregulated by 1,25(OH)2‐vitamin D3 compared to vehicle (Figure 2c). In addition, expression of Fgf23 in the tibia was increased by 1,25(OH)2‐vitamin D3 (Figure 2d).

Figure 2

Memo1 transcripts were not changed by 1,25(OH)2‐vitamin D3 treatment in wild‐type mice. Memo1 transcripts were assessed in kidney (a) and tibia (b) and were indifferent 6 hr after 1,25(OH)2‐vitamin D3 injection, whereas renal Cyp24a1 transcripts were over 10‐fold increased in comparison to vehicle control (c). Fgf23 gene expression in tibia was increased by 1,25(OH)2‐vitamin D3 compared to vehicle (d). n = 5 per condition, *p < .05 (Mann–Whitney U test)

Memo1 is not regulated by dietary calcium

Next, we determined the effect of varying dietary calcium content for 7 days on Memo1 gene expression. RNA was obtained from a previous experiment performed in our lab (366). In the kidney (Figure 3a) and in the tibia (Figure 3b) of these mice exposed to three different calcium‐containing diets (0.17%, 0.82%, and 1.69%), no change in Memo1 gene expression was observed. A 2.5‐fold increase in gene expression of Casr encoding the calcium‐sensing receptor in the bone upon dietary calcium restriction serves as an experimental control for the dietary intervention and was reported for the samples we used in (O'Seaghdha et al., 2013).

Figure 3

Memo1 transcript levels were not influenced by varying dietary calcium contents. Memo1 transcripts in kidney (a) and tibia (b) were not significantly affected by different dietary calcium intakes; n = 5 per diet (Kruskal–Wallis tests). Calcium‐sensing receptor increased 2.5‐fold in bone of mice on 0.17% calcium diet published in (O'Seaghdha et al., 2013) using the same samples serves as an experimental control

Memo1 is not regulated by dietary phosphate

We investigated the influence of different systemic phosphate loads on Memo1 expression. We showed that different dietary phosphate contents (0.2%, 0.8%, 1.5%) did not significantly affect Memo1 transcript levels in kidney (Figure 4a) or in the tibia (Figure 4b). As an experimental control gene we used renal transcripts of Slc34a3 encoding sodium‐dependent phosphate transporter type 2c (NaPi2c). Renal Slc34a3 was increased, as expected, under low phosphate and decreased under high phosphate diets (Figure 4c).

Figure 4

Memo1 transcript levels were not affected by varying dietary phosphate contents. Memo1 transcripts abundance in kidney (a) and in tibia (b) were not significantly changed by dietary phosphate contents. Neither bone Memo1 (b). Renal Slc34a3 transcripts were used as an experimental control gene and were affected by dietary phosphate content (c). N = 5 per condition; *p < .05 (Kruskal–Wallis test with Dunn's multiple comparisons correction)

Memo1 unchanged by PTH

To determine the effect of PTH on Memo1, human PTH fragments 1–34 were subcutaneously injected to wild‐type mice, and the animals were euthanized after 2 hr. Memo1 gene expression in kidney (Figure 5a) or in tibia (Figure 5b) remained unchanged upon PTH treatment. Transcripts of Cyp27b1, the gene coding for the renal vitamin D activating enzyme cytochrome P450 27b1 were increased upon PTH compared to NaCl 0.9%‐treated controls (Figure 5c).

Figure 5

Memo1 gene expression remained unchanged upon PTH treatment in wild‐type mice. Human PTH1‐34 or NaCl 0.9% vehicle was injected 2 hr prior to dissection and transcripts of Memo1 in kidney (a) and tibia (b) were unchanged between experimental conditions. Transcripts of Cyp27b1 (c) were increased by PTH1‐34. n = 6 per condition; ns, not significant; *p < .05 (Mann–Whitney U test). PTH, parathyroid hormone

Memo1 unchanged by estradiol

As sex hormones exert effects on both renal calcium transport proteins (van Abel et al., 2002) and FGF23 in the bone (Carrillo‐Lopez et al., 2009), we tested if Memo1 is a target gene induced by estradiol. We subcutaneously injected 17β‐estradiol once daily over 5 days. This induced the expression of the control gene F3 encoding coagulation factor III (Figure 6c), but gene expression of Memo1 in the kidney (Figure 6a) and in the bone (Figure 6b) both remained unchanged compared to mice injected with vehicle.

Figure 6

Memo1 transcripts remained unchanged upon 17β‐estradiol treatment. Memo1 transcripts assessed by qPCR in kidney (a) and tibia (b) were unchanged after five daily subcutaneous injections of 17β‐estradiol compared to vehicle. Renal gene expression of tissue factor F3 was increased by 17β‐estradiol (c). n = 4 estradiol and n = 5 control condition. ns, not significant; *p < .05 (Mann–Whitney U test)

To summarize, we found a small but significant decrease in Memo1 expression upon 1,25(OH)2‐vitamin D3 exposure in vitro, but we failed to detect any major regulation of Memo1 transcript abundance upon mineral load or calciotropic hormone treatment in vivo.

DISCUSSION

MEMO1 is expressed in the kidney where it plays an intrarenal role in the regulation of calcium transporters (Moor, Haenzi, et al., 2018). In the bone, MEMO1 is expressed in all cell types (Moor, Ramakrishnan, et al., 2018), but its precise bone‐specific function remains elusive.

Here we tested the hypothesis whether Memo1 is regulated by key players in mineral homeostasis such as calciotropic hormones or dietary calcium or phosphate. As a readout, we chose Memo1 gene expression in an osteocyte‐like cell line, and in bone and kidney tissues. For each intervention, an experimental control gene was assessed and revealed effects similar as shown before by others.

We observed that Memo1 gene expression was diminished in the osteocyte‐like cells upon 1,25(OH)2‐vitamin D3 treatment. However, in bone and tissues, we failed to detect any effect on Memo1 by all interventions that we studied. This shows a major difference between Memo1 and the most studied contributors to mineral homeostasis. As examples in the kidney, Type II sodium‐dependent phosphate cotransporters are regulated by dietary phosphate supply (Bourgeois et al., 2013) and gene expression of Trpv5 encoding a renal calcium transport protein is controlled by 1,25(OH)2‐vitamin D3 (Hoenderop et al., 2001). As examples in the bone, Fgf23 expression is stimulated by 1,25(OH)2‐vitamin D3 (Liu et al., 2006) or PTH (Kawata et al., 2007), while dietary phosphate restriction or renal phosphate‐wasting disorders reduce Fgf23 expression (Ansermet et al., 2017; Schlingmann et al., 2016; Vervloet et al., 2011). Even intravenous calcium loading in rats increased Fgf23 expression in bone and hormone concentrations in the serum (Shikida et al., 2018).

This study contains some limitations: The current interventions were confined to the analysis of gene expression, but we did not directly assess Memo1 promoter activity using a reporter construct. Such an approach would more sensitively discriminate and would allow validating putative response elements in the Memo1 promoter. In addition, our in silico analysis of the presumed promoter sequence did not allow base mismatches compared to known response elements. However, we argue that a physiologically relevant the regulation of Memo1 gene expression, if present, should have been visible using the experimental approaches that were undertaken.

Further, we have assessed a single but physiologically reasonable time point, and only a narrow selection of tissues and cells. In addition to bone and kidney, the intestine would be another major turnover place for minerals. Memo1 expression and potential regulation in healthy intestine have not been investigated so far. In colorectal cancer cells Memo1 promoter activity is increased in response to the transcription factors Aryl hydrocarbon receptor/ Aryl hydrocarbon receptor nuclear‐translocator complex, indicating some intestinal disease relevance (Bogoevska et al., 2017).

Another limitation is the fact that we investigated only mice of male sex as to simplify experimental planning, reduction of mice numbers used, and as to reproduce the hormone injection protocols in the cited references, including the estradiol injection protocol (van Abel et al., 2002). Future experiments should be conducted with both sexes independently to allow the detection of sex‐specific effects.

Finally, as Memo is a redox enzyme with incompletely understood reaction partners and substrates (MacDonald et al., 2014), an assessment of posttranslational regulation such as by the oxidative modification of MEMO1 protein, subcellular localization, or changes in its putative enzymatic activity may be helpful to investigate a regulation of Memo1.

To conclude, besides a minor effect in bone cells stimulated with 1,25(OH)2‐vitamin D3, we did not detect a major regulation of Memo1 gene expression upon minerals and calciotropic stimuli in bone and kidney, two organs relevant for mineral homeostasis. Further studies inquiring the regulation of this and similar genes may contribute to the understanding of the regulation of mineral homeostasis in health and renal and bone diseases.

CONFLICT OF INTEREST

The authors declare that no conflict of interest exists.

AUTHOR CONTRIBUTIONS

OB conceived the project. MBM and OB participated in the experimental design. MBM performed the experiments. MBM and OB participated in the data analysis and interpretation. MBM wrote the manuscript. All authors critically read and commented on the manuscript and agreed to the manuscript submission.

Fig S1

Click here for additional data file.