Hyponatremia elicits gene expression changes driving osteoclast differentiation and functions

Abstract

Growing evidence indicates that chronic hyponatremia represents a significant risk for bone loss, osteoporosis, and fractures in our aging population. Our prior studies on a rat model of the syndrome of inappropriate antidiuretic hormone secretion indicated that chronic hyponatremia causes osteoporosis by increasing osteoclastic bone resorption, thereby liberating stored sodium from bone. Moreover, studies in RAW264.7 pre-osteoclastic cells showed increased osteoclast formation and resorptive activity in response to low extracellular fluid sodium ion concentration (low [Na⁺]). These studies implicated a direct stimulatory effect of low [Na⁺] rather than the low osmolality on cultured osteoclastic cells. In the present cellular studies, we explored gene expression changes triggered by low [Na⁺] using RNA sequencing and gene ontology analysis. Results were confirmed by mouse whole genome microarray, and quantitative RT-PCR. Findings confirmed gene expression changes supporting osteoclast growth and differentiation through stimulation of receptor activator of nuclear factor kappa-B ligand (RANKL), and PI3K/Akt pathways, and revealed additional pathways. New findings on low [Na⁺]-induced upregulation of lysosomal genes, mitochondrial energy production, MMP-9 expression, and osteoclast motility have supported the significance of osteoclast transcriptomic responses. Functional assays demonstrated that RANL and low [Na⁺] independently enhance osteoclast functions. Understanding the molecular mechanisms of hyponatremia-induced osteoporosis provides the basis for future studies identifying sodium-sensing mechanisms in osteoclasts, and potentially other bone cells, and developing strategies for treatment of bone fragility in the vulnerable aging population most affected by both chronic hyponatremia and osteoporosis.

Issue sections: Signaling Pathways; Parathyroid, Bone, and Mineral Metabolism.

1. Introduction

Systemic sodium ion ([Na⁺]) and osmotic homeostasis are maintained within narrow limits, defined as serum [Na⁺] concentrations between 135 and 145 mmol/L and plasma osmolality between 280 and 300 mOsm/kg H₂O in humans and other mammals. Chronic hyponatremia ([Na⁺]) <135 mmol/l) is frequently caused by medication use (e. g., diuretics, antidepressants, and seizure medications), and chronic comorbid conditions (e.g., liver, kidney or heart disease). In addition, at least one third of cases are due to the syndrome of inappropriate antidiuretic hormone secretion (SIADH), caused by inappropriate arginine vasopressin (AVP) secretion (Cowen et al., 2013). This electrolyte abnormality is a significant public health problem due to its high prevalence in elderly patients with associated morbidity and mortality. Although hyponatremia is often considered to be asymptomatic, accumulating evidence suggest a strong association with increased all-cause mortality across multiple diseases (Verbalis et al., 2013). Chronic hyponatremia also increases falls and fractures in elderly patients already at high risk for osteoporosis due to other factors.

Growing evidence from clinical studies by several independent groups demonstrated that hyponatremia is associated with bone loss, osteoporosis, and increased bone fragility (Verbalis et al., 2010; Barsony et al., 2011, Barsony et al., 2012; Usala et al., 2015; Hoorn et al., 2011; Jamal et al., 2015; Tolouian et al., 2012; Upala and Sanguankeo, 2016; Ayus et al., 2017). Explanation for these findings come from past and ongoing translational studies in experimental animals and cultured cells (Barsony et al., 2012; Tamma et al., 2013; Fibbi et al., 2016). Our bone histomorphometric analysis utilizing a rat model of SIADH confirmed progressive bone loss associated with combined hyponatremia and hypoosmolality, and revealed 5–10-fold increased osteoclast numbers in excised femurs, tibias, and vertebrae (Verbalis et al., 2010). Because analysis of blood samples revealed no significant metabolic or hormonal change that would account for the increased bone resorption (Verbalis et al., 2010; Barsony et al., 2012), these studies therefore suggested that the observed bone loss was due to direct effect of low [Na⁺] on bone.

Sodium ion (Na⁺) is essential to life and cell integrity, serving to maintain the electrolyte balance inside and outside of cells, conductance of electrical impulses, and cell volume. Consequently, Na⁺ is highly conserved in a diverse array of species from prokaryotes to eukaryotes, including plants, fungi, and animals. In humans, one third of the body’s Na⁺ content is stored in the bone matrix along with calcium and phosphorus, which is released during osteoclastic resorption during growth and prolonged dietary sodium depletion (Barsony et al., 2012; Edelman et al., 1954). The bone matrix is dissolve by osteoclasts, which resorb the inorganic mineral through acidification of the local bone microenvironment and degrade the matrix collagen by secreting proteases. Early radioisotope studies indicated that bone resorption is necessary for liberation of sodium salt from bone. The mechanisms that initiate osteoclastic bone resorption in response to sodium deprivation have not been previously elucidated, but the activation of osteoclastic bone resorption in response to calcium deprivation and various other stimuli has been studied extensively (Dvorak and Riccardi, 2004). Generally, osteoclast activation is initiated by activation of cell membrane sensors: ion channels, membrane receptors for tumor necrosis factor α (TNF), receptor activator for nuclear factor kappa B (RANK) ligand (RANKL), macrophage colony stimulating factor (M-CSF) and G-protein coupled receptors (i.e., PTH or PTH-related protein) (Teitelbaum, 2007; Touaitahuata et al., 2014; Supanchart and Kornak, 2008). Multiple signaling cascades are then activated (Roodman, 2006), including the phosphatidyleinositol-3-kinase (PI3-kinase) pathway, coupled to protein kinase B (Akt), and the mammalian target of rapamycin (mTOR) survival pathway. This PI3-kinase/Akt/mTOR pathway plays a central role in osteoclast differentiation, leading to energy (ATP) production and downstream regulation of osteoclast-specific gene expression necessary for bone resorbing functions (Tiedemann et al., 2017; Golden and Insogna, 2004; Lacombe et al., 2013). Additionally, the roles of the transcription factors NFATc1 and MYC (Lorenzo, 2017) and the lysosomal/mitochondrial signaling pathway (Inpanathan and Botelho, 2019) have also been highlighted.

Activated macrophages and mature osteoclasts secrete matrix metalloproteinase-9 (MMP-9), a zinc dependent endopeptidase. MMP-9 in its active form is responsible for bone matrix degradation and is necessary for osteoclastic cell migration (Ishibashi et al., 2006). The inactive pro-enzyme is a 92 kDa protein (proMMP-9), which is cleaved by the lysosomal cathepsin K enzyme that requires an acidic environment for optimal activation of MMP-9 (Christensen and Shastri, 2015). The 82 kDa activated MMP-9 is a gelatinase, involved in bone matrix degradation. RANKL differentially induces MMP-9 gene expression through p38 and ERK signaling pathways during osteoclast differentiation from precursors (Sundaram et al., 2007; Sabry et al., 2021). Following osteoclast activation, changes in MMP-9 expression have been detected by RT-PCR, gelatin zymography, Western blot analysis of MMP-9 precursor and cleaved protein expression, and resorption assays on bone matrix surfaces (Marroncini et al., 2021; Zhao et al., 2021). We have previously shown that incubation with low extracellular [Na⁺] activates osteoclast differentiation and bone resorption, synergistically with RANKL (Barsony et al., 2011).

Use of mRNA sequencing and network-based analysis of global transcriptomic profiling have recently provided new insights into the regulation of osteoclast formation and activation by micronutrients in cultured monocytic cells (Hanel et al., 2020). Novel osteoclastogenic signaling pathways have also been discovered by recent reports (Erkhembaatar et al., 2017; Mitrofan et al., 2010; Liu et al., 2017).

Here we use transcriptome analysis using next generation mRNA sequencing to explore how RAW264.7 murine monocytic osteoclastic cells in culture respond to low extracellular [Na⁺] for 24 h. We previously found that low [Na⁺] directly stimulates osteoclastogenesis and osteoclastic resorption in cultured osteoclastic cells (Barsony et al., 2011). We substantiated the findings from mRNA sequencing and network based analysis by mouse whole genome microarray and qRT-PCR, and correlated transcriptomic assay results with functional assays. Combined, our results reveal the magnitude and diversity of differentially affected genes, and functional analyses demonstrate activation of lysosomal/mitochondrial pathways that transmit the osteoclastogenic response to low [Na⁺].

2. Material and methods

2.1. Cell culture

Low passage number murine monocytic RAW264.7 cells were obtained from the American Type Culture Collection, Manassas, VA (ATCC). Bone marrow was prepared from 4-month-old male 129/Sv mice for primary culture of osteoclasts as described. Cells were grown in Minimal Essential Media Alpha with glucose 4.5 g/L (αMEM; Invitrogen, Carlsbad, CA) with 10% fetal bovine serum (ATCC), 100 μg/ml penicillin/streptomycin (Invitrogen) and 2 mM glutamine (normal medium; [Na⁺] = 140 mmol/l). To generate media with low [Na⁺] = 120 mmol/l or [Na⁺] = 128 mmol/l concentrations, the normal [Na⁺] medium was diluted with a mixture of water, minerals (except sodium), amino acid concentrate, fetal bovine serum, and vitamins (Mediatech Inc., Herndon, VA) to match the normal medium, while maintaining normal osmolality = 290 mOsm/L via the addition of mannitol. Sodium and osmolality of media was measured before application to cells using a sodium analyzer (Coulter ELISE, from Beckman, Fullerton, CA) and an osmometer (Model 3900, Advanced Instruments, Inc., Norwood, MA).

2.2. Gene expression analysis

2.2.1. Groups and RNA extraction

RAW264.7 cells were incubated in cell culture media with normal [Na⁺] or low [Na⁺] with cytokines for 24h with at least 3 biological replicates in each group. Total RNA was extracted using PureLink kit (Invitrogen, Carlsbad, CA) according to manufacturer’s instructions, and purified using the Rneasy Mini Kit (Qiagen, Germantown, MD). Agarose gel electrophoresis was used to check the integrity of total RNA.

2.2.2. Next generation mRNA sequencing and analysis of gene expression

All mRNA samples were sent for sequencing and data analysis to Arraystar Inc. (Rockville, MD). From each sample, 2 μg total RNA was used to prepare the sequencing library using KAPA Stranded RNA-Seq Library Prep Kit (Illumina), which incorporates dUTP into the second cDNA strand and renders the RNA-Seq library strand-specific. The completed libraries were qualified with Agilent 2100 Bioanalyzer and quantified by absolute quantification qPCR method. To sequence the libraries on the Illumina NovaSeq 6000 instrument, the barcoded libraries were mixed, denatured to single stranded DNA in NaOH, captured on Illumina flow cell, amplified in situ, and subsequently sequenced for 150 cycles for both ends on Illumina NovaSeq 6000 instrument. Image analysis and base calling were performed using Solexa pipeline (Off-Line Base Caller software, v1.8) and sequence quality was examined using the FastQC software. The trimmed reads (trimmed 5′, 3′-adaptor bases using Cutadapt) were aligned to reference genome using Hisat2 software. The transcript abundances for each sample was estimated with StringTie, The differentially expressed genes and transcripts were filtered using R package Ballgown. Principle Component Analysis (PCA) and correlation analysis were based on gene expression level, hierarchical clustering, gene ontology, and pathway analysis. Scatter plots and volcano plots were performed with the differentially expressed genes in R, Python or shell environment for statistical computing and graphics.

2.2.3. Microarray analysis of gene expression

RNA samples were submitted to the National Institute of Diabetes and Digestive Diseases Genomic Core Facility for global gene expression analysis utilizing Affymetrix GeneChip Mouse Genome 430 2.0 whole genome arrays. These arrays contain 45,101 probe sets corresponding to known genes and expressed sequence tags. RNA samples were validated and processed according to published protocol. 5 μg RNA from each sample was converted into biotin-labeled cDNA and hybridized to arrays using the Affymetrix Gene Chip expression 3’ amplification reagents in a two-cycle cDNA synthesis kit (Affymetrix). Microarray data were normalized and analyzed using the Microarray Analysis Suite 5.0 (Affymetrix). Mean signal intensities and the p-value of significance were calculated for each probe sets and data sets were filtered with signal detection threshold of 50 and p < 0.05 after Bonferroni correction. The fold change was determined by dividing the normalized gene expression in the low [Na⁺] sample with the normalized gene expression in normal [Na⁺] sample by analysis of variance using the Partek Pro software (Partek). Positive values indicate up-regulation and negative values indicate down-regulation of mRNA expression. Array data were analyzed for significantly affected pathways using the Ingenuity Pathway analysis software application. The top significantly affected pathways are indicated by log p-values.

2.2.4. Quantitative RT-PCR validation of gene expression

The mRNA was prepared from RAW264.7 cells exposed to normal [Na⁺] or low [Na⁺] differentiation media for 24 h. Preliminary experiments also included 48 h exposures. Relative gene expression was determined using quantitative RT-PCR using reagents, probes and primers from Qiagen. The 18S ribosomal RNA was used for normalization.

2.3. Osteoclast functional assays

2.3.1. Lysosomal acidity changes

To assess localization and distribution of acidic lysosomes, RAW264.7 derived osteoclasts were subcultured into coverslip chamber slides (Nunc) (10⁵/well). The next day, cells were loaded with the pH sensitive dye LysoSensor Green DND-189 (ThermoFisher; 443/505) at 0.5 μl/ml in normal [Na⁺] media for 30 min and then media was changed to fresh normal [Na⁺] or low [Na⁺] media without dye for additional 30 min. Images were taken with a Zeiss LSM 510 laser scanning confocal microscope using excitation of 488 nm and emission of 514 nm from an argon laser. At least 20 images were taken from each condition, with 3 repeat experiments.

For fluorometric lysosomal acidity measurements, RAW264.7 derived osteoclasts were subcultured into 96-well plates (10⁵/well). For the assay, media was changed to HANKS balanced salt solution with normal [Na⁺] =140 mmol/l or low [Na⁺] =120 mmol/l in 6 wells each. LysoSensor Green DND-189 was added to each well to reaching a final concentration of 0.5 μl/ml. Incorporation of dye and subsequent acidity change was monitored over 2 h (3, 5, 10, 15, 30, 60, 120 min) as relative fluorescence using the excitation filter at 440 nm and emission filter at 520 nm. Each experiment was repeated 3 times.

The effects of RANKL exposure on lysosomal acidity were tested in differentiated RAW264.7 derived osteoclasts, subcultured into 96-well plates. First, cytokines were withdrawn for 24 h. Cells were then exposed to RANKL (50 ng/mL) for 24h in normal [Na⁺] or low [Na⁺] HANKS balanced salt solution. Finally, cells were loaded with LysoSensor Green DND-189 for 30 min, and washed before fluorometric lysosomal acidity measurements.

2.3.2. Cytotoxicity and mitochondrial ATP production

We evaluated how low [Na⁺] influences cell death and energy balance in RAW264.7 cells. Cells were subcultured into 96-well plates for 24 h and then exposed to media with normal [Na⁺] = 140 mmol/l (6 samples) or with low [Na⁺] = 120 mmol/l (6 samples) with cytokines. Cytotoxicity and mitochondrial ATP production were both measured using a dedicated assay kit “ToxGlo” from Promega Corp. (Madison, WI USA) according to manufacturer’s instructions.

Cytotoxicity assay detected damage to cell membrane integrity by measuring “dead cell” protease activity with a fluorescent marker bis-AAF-R110 substrate. Treatment with digitonin (80 μg/ml) for 90 min served as positive control. Mitochondrial ATP production was measured after 24h in normal or low [Na⁺] differentiation medium. Then cells were treated for 30 min at 37 °C in serum free media with galactose instead of glucose while maintaining the same [Na⁺] in the media. This galactose addition suppresses energy production by glycolysis while favoring ATP production by oxidative phosphorylation. Cells were lysed and ATP content was measured in the lysate with the addition of a mix containing ATPase inhibitor, luciferin, and thermostable luciferase to assure luminescence value proportional to the ATP content. Data were normalized to protein content of the lysate.

2.3.3. MMP-9 expression and cell migration

MMP-9 expression was evaluated by Western blot analysis in lysates from RAW264.7 cells. Cells were differentiated to mature osteoclast over 6 days in media with cytokines, then cytokines were withdrawn for 24 h, followed by incubation with or without RANKL (50 Ng/ml) in either normal [Na⁺] or low [Na⁺] for 24h and 48h. Lysates were prepared in RIPA buffer+ 20 μl protease inhibitor cocktail (Thermo, #78425) and for electrophoresis the lysate samples were adjusted for protein content of 18 μg/lane. Primary and secondary antibodies were from Abcam (Woburn, MA). We used primary rabbit recombinant multiclonal IgG against MMP-9 Cat# ab283575 1:1000 dilution (0.615 μg/mL) at 4 °C overnight followed by incubation with secondary anti-rabbit IgG Cat# ab97051 in 1:20,000 dilution. Findings were confirmed with another Western blot analysis using antibodies from R&D Systems Inc. (Minneapolis, MN), including primary goat polyclonal antibody against MMP-9 Cat# AF909 at 20 μg/mL followed by secondary anti-goat IgG Cat# HAF019 at 1:1000 dilution.

To test low [Na⁺] effects on cell migration, RAW264.7 cells were seeded into 96-well Boyden chamber assay kit inserts (Cultrex 96-well Cell Migration Kit, Trevigen Inc., Gaithersburg, MD) at the density of 25,000 cells/well. During a 24h incubation at 37 °C, cells migrated through the insert membrane into the bottom of the wells. In each of the experiments, six wells were incubated with normal [Na⁺] differentiation media and other six wells with low [Na⁺] differentiation media. Standard curves were generated to correlate calcein-AM fluorescence intensity with cell number. Fluorescent indicator was added after the removal of the inserts. The assay protocol used was according to the manufacturer’s instructions. Data are expressed as percent cells migrated compared to the initial number of cells placed onto the insert.

2.3.4. Statistical analysis of bioassays

Statistical differences were examined using parametric or non-parametric t-tests or one-way ANOVA with Dunn’s multiple comparisons test as appropriate using SigmaStat software. The statistical test used for each dataset is noted in the figure legends. An adjusted p value < 0.05 was considered statistically significant. Data are expressed as mean ± SEM.

3. Results

3.1. Low [Na⁺] elicits gene expression changes in RAW264.7 osteoclastogenic cells

3.1.1. Results of mRNA sequence analysis

The high quality of our sequence analysis is shown by the correlation plot of expression profiling (Fig. 1A) and volcano plot of the genes (Fig. 1B) for low [Na⁺] and control groups.

Fig. 1. Transcriptomic analysis shows the effects of 24h exposure to normal versus low sodium concentration in the extracellular medium on differential gene expression in RAW264.7 osteoclastic cells.

A: Scatter plot shows close correlation of differentially expressed genes identified by RNA-sequencing of samples from RAW264.7 cells in culture exposed to low ECF [Na⁺] (LS, 120 mmol/l) compared to expression from cells in normal ECF [Na⁺] (NS, 140 mmol/l). Red data points represent upregulated genes and green data points represent downregulated genes, whereas black data points represent genes unaffected by [Na⁺]. Gene expression was calculated by comparing FPKM (fragments per kilo base of exon per million mapped fragments) values in LS and NS samples, obtained by RNA-sequencing. The figure is based on the Poisson distribution method.

B. Volcano plot of differentially expressed genes identified genes significantly over- or under-expressed in samples from RAW264.7 osteoclastic cells exposed to low ECF [Na⁺] (120 mmol/l) compared with the expression of the same gene in samples from cells in normal ECF [Na⁺] (140 mmol/l) obtained by RNA-sequencing. Black dots below the horizontal line represent genes affected below the significance level of 0.05. The vertical lines separates genes induced or suppressed by low ECF [Na⁺] by less than 1.5-fold (non-significantly affected). Above the horizontal line, red dots represent genes induced by low [Na⁺] exposure at or above 1.5-fold (red dots) and green dots represent significantly down-regulated by low ECF [Na⁺].

3.1.1.1. Differentially expressed genes (DEGs).

With the threshold of adjusted p < 0.05, false discovery rate (FDR) below 0.1, and log fold change cutoff of 1.5 induction or inhibition, DEGs included 1812 genes significantly upregulated by 24 h exposure of cells to low [Na⁺] media compared to expression in cells cultured in normal media (control). The same low [Na⁺] exposure selectively downregulated 1760 genes, whereas 6939 genes were unaffected. A detailed list of significantly affected genes can be found at Supplemented Material 1.

Among the DEGs, several osteoclast-specific marker genes were markedly upregulated, including genes associated with bone resorption, osteoclast polarization and attachment, and motility. Upregulation of sodium/hydrogen exchanger (Atp6v0d2) and vacuolar proton transporting 4 (carbonic anhydrase 2) promote acid secretion, and tartrate-resistant acid phosphatase (Acp5), matrix metalloproteinase-9 (MMP9), and cathepsin K (Ctsk) promote matrix protein degradation. Among the marker genes, the upregulation of tyrosine kinase Src gene promotes polarization and attachment and the upregulation of Ankyrin gene promotes motility (Table 1). These osteoclast functions require differentiation of monocytic cells into osteoclasts; a process likely promoted by the upregulated genes linked to osteoclast differentiation listed in Table 1. Downregulation of several marker genes also contribute to osteoclastic bone resorption by reducing inhibitory signals, such as the top affected genes heparanase, CD200 receptor, Igf1, tetraspanin10, Atf3, and GSTm1 (Table 1).

Table 1. List of selected differentially expressed osteoclast specific genes from RNA-sequencing experiments comparing low [Na⁺] versus normal [Na⁺] effects.

Samples were from RAW264.7 osteoclastic cells exposed for 24h to media with low [Na⁺] (120 mmol/l) or normal [Na⁺] (140 mmol/l). P < 0.05 was considered statistically significant.

Gene name	Protein	Protein function	Fold change	p-value	FDR q-value	FPKM of gene in samples		FPKM of gene in samples
						Mean low [Na+] mean	S.D. low [Na+]	Mean normal [Na+]	S.D. normal [Na+]
	Top selectively up regulated genes
	osteoclast specific functions
Acp5	tartrate resistent acid phosphatase	matrix degradation	119	1.14E-09	1.42072172426566E-05	956.869	20.332	7.048	0.107
Ctsk	cathepsin K	matrix degradation	41	8.06E-09	3.23715251716872E-05	1091.835	23.539	25.497	0.615
Car2	carbonic anhydrase 2	pH regulation	15.1	1.95E-08	3.23715251716872E-05	191.299	2.109	11.110	0.335
Ank	ankyrin	Na ion transport, motility	14.7	3.3321053150015E-07	9.37992336989715E-05	549.470	31.474	36.470	0.330
Mmp9	Matrix metalloproteinase 9	matrix degradation	5.3	1.05E-06	0.000125719239559815	226.789	10.066	41.638	0.806
Src	src kinase	polarization and attachment	2.7	0.0114380736211345	0.0236459993943777	48.751	0.531	23.527	6.642
Atp6v0d2	vacuolar ATPase	acid secretion	1.8	7.84839173536556E-05	0.000861393723299428	276.710	11.962	155.120	3.812
	osteoclast regulators
Dmpk	effector of Rho GTPase	osteoclast formation	13.7	1.61876136162142E-07	6.99679395159155E-05	18.940	0.926	0.457	0.010
ocstamp	osteoclast stimulatory transmembrane protein	osteoclast differentiation	13.1	1.12945768959705E-07	6.5264752779995E-05	16.122	0.211	0.307	0.054
Gsttl	glutathione-S-transferase theta 1	osteoclast formation	11.7	1.12896815884778E-05	0.000333377019195173	23.545	1.471	1.111	0.237
Traf1	receptor associated factor 1	osteoclast formation	5.3	2.89036713230217E-07	9.37992336989715E-05	14.308	0.357	1.877	0.074
Slc9b2	Na+/H+ antiporter	osteoclast differentiation	11.2	2.6482004722439E-07	8.94733894689757E-05	10.568	0.525	0.026	0.020
Oscar	osteoclast associated receptor	osteoclast differntiation	9.6	8.70E-07	0.000119460779583085	9.522	0.504	1.033	0.102
Myc	c-Myc proto-Oncogen	osteoclast differntiation	8.2	3.89079166107642E-07	9.40352645726255E-05	55.079	1.686	5.872	0.253
Fos	c-Fos proto-Oncogen	osteoclast proliferation and survival	6.8	1.08096280249104E-06	0.000127482226357929	82.766	2.079	11.556	0.609
Nfatc1	nuclear factor of activated T cells	osteoclast differntiation	5.9	1.14730317208878E-05	0.000333377019195173	48.814	4.073	7.336	0.397
dcstamp	dendritic cell-specific transmembrane protein	osteoclast differntiation	2.4	0.019	0.03525231098499	18.052	1.065	7.483	7.004
Siglec15	sialic acid binding lectin	osteoclast differntiation	2	0.018	0.0334524480315347	5.618	1.392	2.247	0.478
cdc26	cell division cycle 26	osteoclast formation	1.9	0.000542406918978777	0.00266953893470618	5.234	0.522	2.293	0.097
Sirt3	sirtuin 3	osteoclast activity	1.6	6.91034552151581E-05	0.00081113830389173	2.212	0.063	1.050	0.065
	Top selectively down regulated genes
Hpse	heparanase	inhibits osteoclastogenesis	−4.9	4.03165376905434E-06	0.000201638691036798	1.341	0.331	66.318	1.006
Cd200r4	CD200 receptor	inhibits osteoclast differentiation	−4.6	1.37942238239708E-05	0.000357775809762266	2.428	0.530	79.541	5.138
Tspan10	tetraspanin10	inhibits osteoclastogenesis	−3.9	2.63749230677846E-07	8.94733894689757E-05	2.270	0.069	49.365	2.571
Atf3	activating transcription factor 3	inhibitor of osteoclast proliferation	−3.6	3.84626621108275E-06	0.000201638691036798	10.542	0.765	137.453	9.880
Gstm1	Glutathione S-Transferase Mu 1	inhibits focal adhesion	−3.4	6.36292020739226E-06	0.00025742027673984	19.035	0.351	214.016	23.478

3.1.1.2. Functional enrichment analysis of differentially expressed genes.

Analysis of most affected canonical KEGG pathways show the expected low [Na⁺] induced enrichment in osteoclast differentiation. Consistent with our previously published data, signaling through the PI3K/MAPK and NFκB pathways are highly affected (Table 2). The enrichment of the actin cytoskeleton pathway with 31 DEGs represent a logical extension to the PI3K/MAPK and NFκB signaling activation. We also reported previously that low [Na⁺] reduced ascorbic acid uptake, which contributed to increased osteoclastogenesis; here this pathway is again implicated (Table 2).

Table 2. Functional enrichment analysis based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways and gene ontology (GO) analysis of differentially expressed genes from RNA-sequencing data.

List of selected pathways characterize the signaling response in RAW264.7 osteoclastic cells exposed for 24h to media with low [Na⁺] (120 mmol/l) compared with expression from cells in normal [Na⁺] (140 mmol/l). Results from RNA-sequencing data revealed that the lysosomal pathway was disproportionately affected by low [Na⁺] exposure, with over 18-fold enrichment. Results also show high impacts on the PI3K/AKT/mTOR pathway, MAP-kinase pathway, and NF-kappa-B pathway, and regulation of mitochondrial energy production, actin cytoskeleton, and ascorbate metabolism. Gene expression drives regulatory changes inducing osteoclast functions including motility, adhesion, acid and proteinase secretion, and autophagy. P-value stands for the fisher exact test value of the term and FDR stands for the false discovery rate of the term, using Benjamini and Hochberg (1995) method.

	p-value	FDR	Enrichment	genes
Kegg pathway term
Lysosome	2.73108242969185E-19	8.05669316759095E-17	18.5636651917606	Ap1g1//Ap3m2//Ap3s1//Ap4e1//Arsa//Arsg//Asah1//Atp6v0a1//Cd164//Cd68//Cln5//Cltc//Ctns//Ctsa//Ctsb//Ctsc//Ctsd//Ctse//Ctsl//Ctso//Ctss//Ctsz//Gaa//Galc//Gba//Gla//Gnptg//Gns//Gusb//Ids//Lamp1//Lamp2//Laptm4a//Laptm4b//Lgmn//Lipa//Litaf//Man2b1//Mcoln1//Naglu//Neu1//Npc1//Pla2g15//Ppt1//Psap//Slc11a1//Sort1//Sumf1//Tpp1
Osteoclast_differentiation	0.000139855166622009	0.00275048494356618	3.85432148500228	Akt3//Fcgr1//Fcgr2b//Fcgr3//Fyn//Gab2//Gm14548//Gm15448//Il1a//Irf9//Jak1//Jun//Lilra6//Lilrb4a//Mapk12//Pik3cb//Pirb//Pparg//Ppp3r1//Sirpa//Sirpb1a//Sirpb1b//Stat2//Tgfb1//Tgfbr2//Tyk2
MAPK_signaling_pathway	0.000233069190293451	0.00404443594920989	3.63251513261368	Akt3//Braf//Cacna1a//Cacna1b//Cd14//Crkl//Ddit3//Dusp1//Dusp16//Dusp3//Dusp4//Dusp6//Dusp7//Fas//Fgf13//Flnc//Gadd45a//Gna12//Gng12//Il1a//Jun//Map3k1//Map3k2//Map3k20//Map3k3//Map4k1//Mapk12//Mknk1//Nf1//Pla2g4a//Ppp3r1//Prkacb//Rapgef2//Rasa1//Rasgrp3//Rps6ka2//Sos1//Sos2//Taok1//Taok3//Tgfb1//Tgfbr2
Phosphatidylinositol_signaling_system	0.000632673874945795	0.00927122293124602	3.19882009852826	Cds2//Dgka//Dgkd//Dgkh//Dgkq//Inpp4a//Inppl1//Itpkb//Itpr2//Mtm1//Mtmr3//Mtmr4//Mtmr6//Pi4ka//Pik3cb//Pikfyve//Pip4k2a//Pip4k2c//Plcg1//Pten
NF-kappa_B_signaling_pathway	0.0014036206756552	0.0128213499059841	2.85275024315196	Atm//Bcl2//Bcl2a1a//Bcl2a1b//Bcl2a1d//Bcl2l1//Ccl4//Cd14//Cd40//Cflar//Cxcl2//Ddx58//Irak1//Malt1//Plcg1//Ptgs2//Ripk1//Tab3//Ticam2//Trim25
Ascorbate_and_aldarate_metabolism	0.0115647164809968	0.0631776178128528	1.93686500986681	Aldh2//Ugdh//Ugt1a1//Ugt1a2//Ugt1a6a//Ugt1a6b//Ugt1a7c
Regulation_of_actin_cytoskeleton	0.0123605023966884	0.0662972401276922	1.90796387684629	Arhgap35//Arhgef7//Braf//Cd14//Cfl2//Crkl//Cyfip2//Diaph1//Dock1//Ezr//Fgf13//Gna12//Gng12//Iqgap1//Itga7//Itgam//Itgax//Itgb1//Msn//Pfn2//Pik3cb//Pikfyve//Pip4k2a//Pip4k2c//Rock2//Slc9a1//Sos1//Sos2//Ssh1//Ssh2//Vav3
Gene Ontology (GO) term biological processes
regulation_of_cell_motility	3.47370660338332E-20	3.81481320263358E-18	19.4592068657716	Abhd6//Acvr1b//Adam15//Adam17//Adam9//Ago2//Aif1//Akt3//Aldoa//Amotl1//Anxa3//App//Atp7a//Atp8a1//Bcl2//Bmpr1a//Bmpr2//Braf//Bst1//C3ar1//C5ar1//Calr//Ccbe1//Ccl4//Cd274//Cd40//Cd74//Cd81//Ceacam1//Col18a1//Cpeb1//Cpne3//Cxcr4//Dab2//Dag1//Ddx58//Diaph1//Dock1//Dock10//Dock4//Ecm1//Evl//Flcn//Fpr2//Gab2//Gas6//Gna12//Gpnmb//Gsk3b//Hdac4//Igf1//Igsf8//Il1a//Il6st//Insr//Iqgap1//Itgb1//Jag1//Jun//Kitl//Lgals3//Lgals9//Lgmn//Lrp1//Macf1//Map3k1//Map3k3//Mia3//Mink1//Mmp12//Mospd2//Mpp1//Msn//Myo1f//Myo5a//Nf1//Nfe2l2//Nipbl//Nod2//Nrp1//Osbpl8//P2rx4//P2ry6//Pdpn//Pfn2//Plcg1//Pld2//Plk2//Plpp3//Plxna1//Plxna2//Plxna3//Plxnc1//Ppard//Pparg//Prcp//Prkce//Prkd2//Prkx//Ptafr//Pten//Ptgs2//Ptk2b//Ptpn23//Ptprc//Rab5a//Rap2a//Rap2b//Rapgef2//Rhob//Rhoc//Rhod//Rhoj//Rock2//Rreb1//S1pr1//Sash1//Sdc3//Sdcbp//Selenok//Sema4c//Slc8a1//Slk//Smad3//Spag9//Spry2//Srgap2//Ssh1//Ssh2//Stap1//Stat3//Stk24//Sun2//Tgfb1//Tgfbr2//Trf//Trp53inp1//Trpv4//Vegfb
cell_death	2.65730168296323E-19	2.57989332958996E-17	18.5755591374194	1600012H06Rik//Aatk//Acvr1b//Aktip//Als2//Anxa6//App//Asah2//Atm//Atn1//Atp7a//Bbc3//Bcl2//Bcl2a1a//Bcl2a1b//Bcl2a1d//Bcl2l1//Birc6//Bmpr2//C5ar1//Cacna1a//Cadm1//Casp2//Cd24a//Cd5l//Cflar//Clu//Csrnp2//Ctsc//Cxcl2//Cyfip2//Dab2//Dapk1//Ddit3//Dnajc10//Dock1//Dram2//Ep300//Ercc6//Ern1//Ero1l//Fas//Fem1b//Fgf13//Fnip2//Fzd5//Gas6//Gclm//Gpr65//Gsk3b//Hcar2//Hip1//Hipk1//Hipk3//Htatip2//Ifi204//Ifi27l2a//Igf1//Ikbke//Il1a//Irak3//Jun//Kank2//Kitl//Lamp1//Lcn2//Lgals1//Lsp1//Magi1//Malt1//Map3k1//Map3k9//Mgea5//Mknk1//Naip2//Naip5//Naip6//Nf1//Nlrc4//Nlrp1b//Ogt//Oxr1//P2rx4//P2rx7//Pacs2//Parp4//Pdcd6ip//Pea15a//Phlda3//Pik3cg//Pmp22//Ppard//Ppp1r15a//Prkcd//Prkd2//Prune2//Pten//Rabep1//Rb1//Rhbdd1//Rhob//Ripk1//Rnf130//Rnf144b//Rock2//Rps6ka2//Rtn3//Rybp//Selenok//Serpinb9//Sgk1//Sgms1//Sgpl1//Sgpp1//Sirt2//Slc9a1//Slk//Smad3//Sort1//Sp110//Stk24//Susd6//Taok1//Tax1bp1//Tctn3//Tgfb1//Tgfbr2//Them4//Tlr3//Tmem173//Tnfrsf10b//Tnfrsf1b//Tnfsf12//Trim35//Trp53bp2//Trp53inp1//Trpm7//Vps35//Zfp385a//Zmat3
regulation_of_GTPase_activity	5.73246345857646E-13	3.39838696539856E-11	12.2416587049869	Acap3//Adap2//Als2//Als2cl//Arfgap3//Arfgef1//Arhgap12//Arhgap17//Arhgap22//Arhgap35//Arhgef7//Asap1//Cav2//Ccl3//Ccl4//Ccl6//Cd40//Chm//Chml//Dock1//Dock10//Dock11//Dock2//Dock4//Evi5//Fam13b//Gpr65//Gpsm1//Gsk3b//Iqgap1//Itgb1//Jun//Myo9a//Myo9b//Nf1//Nrp1//Pkp4//Plxna1//Plxna2//Plxna3//Plxnc1//Ptk2b//Rab3gap2//Rabep1//Ralgapb//Ranbp2//Rap1gds1//Rapgef2//Rasa1//Rasgrp3//Rgs1//Rgs18//Rictor//Rin2//S1pr1//Sbf1//Snx13//Snx18//Spry2//Srgap2//Stard8//Stxbp5//Syngap1//Tbc1d16//Tbc1d2//Tbc1d20//Tbc1d22a//Tbc1d4//Tbc1d8b//Tbc1d9//Tbc1d9b//Tsc1//Vav3
regulation_of_cell_adhesion	8.87547800076692E-13	5.17015888062066E-11	12.0518082481314	Adam15//Adam9//Aif1//Arg2//Ass1//Bcl2//Braf//Bst1//Calr//Cask//Cblb//Cd24a//Cd274//Cd276//Cd36//Cd74//Ceacam1//Crkl//Dab2//Dag1//Dock1//Dusp3//Emilin2//Flcn//Flot2//Fzd7//Gpam//Gpnmb//Gsk3b//H2-M3//H2-T23//Hfe//Hlx//Igf1//Il6st//Il7r//Iqgap1//Itgb1//Itpkb//Jag1//Kifap3//L1cam//Lgals1//Lgals3//Lgals9//Macf1//Malt1//Mia3//Mink1//Myo1f//Nck2//Nf1//Nfkbiz//Nrp1//Pag1//Pdpn//Peak1//Pik3cb//Pik3r6//Pkd1//Pld2//Plpp3//Plxna1//Plxna2//Plxna3//Plxnc1//Prkcd//Prkce//Prkd2//Prkx//Ptafr//Pten//Ptk2b//Ptpn23//Ptprc//Rasa1//Rc3h1//Rc3h2//Rhod//Rreb1//S1pr1//Selenok//Sirpa//Slk//Smad3//Spp1//Stat5b//Stx3//Tarm1//Tgfb1//Tgfbr2//Tnfsf9//Trpv4//Tsc1//Utrn//Vav3//Zmiz1
regulation_of_cytoskeleton_organization	7.53047540398149E-13	4.42514515186597E-11	12.1231776056308	Add3//Akap13//Arfgef1//Arhgap17//Arhgap35//Atf5//Avil//Braf//Bst1//Bst2//Camsap1//Cep192//Cfl2//Chmp1b//Clip1//Cltc//Ctnnb1//Ctsl//Dctn1//Diaph1//Dst//Dync1h1//Ep300//Eps8//Evl//Fgf13//Gda//Gmfg//Gpr65//Gsk3b//Il1a//Kank2//Lats1//Lima1//Lrp1//Map3k1//Mtpn//Mtss1//Mycbp2//Myo1f//Nck2//Nin//Nrp1//Pdcd6ip//Pdlim4//Pfn2//Pkd1//Plek//Plk2//Plxna3//Ppfia1//Prkcd//Prkce//Psrc1//Ptk2b//Rasa1//Rhob//Rhoc//Rhod//Rhoj//Rhoq//Rhov//Rictor//Rock2//S1pr1//Sgk1//Sh3pxd2b//Smad3//Spire1//Spire2//Sptan1//Sptbn1//Ssh1//Ssh2//Stag1//Stap1//Stau2//Taok1//Tgfb1//Tmod1//Togaram1//Tpr//Trpv4//Tsc1//Tubb4a
positive_regulation_of_secretion	1.80558349995658E-09	5.92921758147506E-08	8.7433824226446	Aacs//Acsl3//Acvr2a//Adam9//Ap1g1//Arhgef7//Atg7//Bcl2l1//Bglap2//Cacna1b//Cadm1//Cadps//Cask//Ccl3//Cd14//Cd274//Cd276//Cd300c2//Cd36//Ceacam1//Clec4e//Dab2//Ddx58//Dtnbp1//Ep300//Exoc2//Ezr//Fcer1g//Fzd5//Gab2//Glul//H2-T23//Hcar2//Hfe//Ifih1//Igf1//Il1a//Il1rl1//Itgam//Kif5b//Lamp1//Lgals3//Lgals9//Lpl//Mcu//Mgea5//Mmp12//Nlgn2//Nod2//Nucb2//P2rx7//Panx1//Pdcd6ip//Pld2//Ppard//Prkce//Ptafr//Ptpn23//Rab5a//Rhbdd1//Sdcbp//Sec24a//Selenok//Spp1//Sptbn1//Stxbp5//Tgfb1//Tlr1//Tlr8//Trpv4//Vamp7//Vps35//Wls
proteolysis	1.30353087110097E-08	3.56422583898179E-07	7.88487867911779	Adam15//Adam17//Adam9//Adamts1//Amz1//Anpep//Aph1b//Aph1c//Asph//Ate1//Atg4c//C1ra//C1rb//Capn1//Capn15//Casp2//Cbl//Ccdc47//Cd5l//Cfb//Cflar//Cln5//Clock//Cndp2//Cpd//Ctsa//Ctsb//Ctsc//Ctsd//Ctse//Ctsl//Ctso//Ctss//Ctsz//Cul9//Cyfip2//Dag1//Ddit3//Derl1//Dnajb9//Dnajc10//Dnajc3//Dpep2//Dpp7//Dpp8//Dtx3l//Edem1//Edem2//Edem3//Erap1//Ero1lb//F10//Fbxl17//Fbxl5//Fbxw4//Foxred2//Ggh//Hectd3//Herc2//Hp//Hsp90b1//Hspa13//Ift172//Kctd17//Kctd21//Kctd6//Lgmn//Lnpep//Malt1//March6//Mbtps2//Mindy1//Mindy2//Mmp12//Mmp19//Ncstn//Nfe2l2//Nhlrc3//Ogt//Os9//P2rx7//Pdcd6ip//Pdia3//Prcp//Psen2//Psma8//Psme4//Ptpn23//Rhbdd1//Rhbdf2//Rnf11//Rnf115//Rnf144b//Rnf213//Scpep1//Sdf2l1//Sel1l//Senp5//Sirt2//Smurf1//Sppl2a//St14//Stt3b//Taf1//Tgfb1//Tollip//Tpp1//Trim25//Trip12//Trp53inp2//Trrap//Ttc3//Uba6//Uba7//Ube4a//Ubr3//Ubr4//Uggt1//Usp18//Usp24//Usp31//Usp32//Usp33//Usp9x//Vcpip1//Vps35//Wdr81//Wfs1//Wwp1
positive_regulation_of_autophagy	1.96916314294321E-07	4.29688074741909E-06	6.70571830156381	Atg7//Dapk1//Flcn//Gba//Gnai3//Gpsm1//Gsk3b//Irgm1//Irgm2//Nod2//Optn//Pip4k2a//Pip4k2c//Plk2//Rab12//Rab3gap2//Smcr8//Smurf1//Tfeb//Tlr9//Trim21//Trp53inp1//Trp53inp2//Tsc1
positive_regulation_of_cell_proliferation	3.57035860223322E-07	7.42789822247215E-06	6.44728816171702	Adam17//Aif1//Akt3//Atf3//Atp7a//Bcl2//Bcl2l1//Birc6//Bmpr1a//Bmpr2//Bst1//C3ar1//C5ar1//Cacul1//Calr//Cav2//Ccng1//Ccpg1//Cd24a//Cd274//Cd276//Cd40//Cd74//Cd81//Ceacam1//Cflar//Clu//Col18a1//Crkl//Ctnnb1//Ctsz//Ecm1//Ern1//Fabp4//Fbxw4//Fndc3b//Foxj2//Fzd7//Gab2//Gas6//Glul//Gnai3//Gpam//Grn//Gsk3b//H2-T23//Hdac4//Hipk1//Hlx//Hmga2//Hpse//Id2//Igf1//Il11ra1//Il6st//Il7r//Insr//Irak1//Itgb1//Jun//Kii3a//Kitl//Kmt2c//Kmt2d//Lgals3//Lgmn//Map3k3//Mmp12//Nampt//Nck2//Nf1//Nlgn2//Nod2//Notch2//Nqo2//Optn//Pggt1b//Pla2g4a//Plcg1//Ppard//Prkd2//Ptafr//Pten//Ptgs2//Ptk2b//Ptprc//Rab5a//Rasa1//Rictor//Rreb1//Runx2//S1pr1//Scg2//Sdcbp//Selenok//Selenon//Slc39a10//Stat3//Stat5b//Stx3//Syne1//Tgfb1//Tgfbr2//Tlr9//Tnfsi9//Trf//Ufl1//Vav3//Vegfb//Zmiz1
vacuolar_acidification	0.00110116049065754	0.008557626597349	2.95814937936846	Atp6v0a1//Cln5//Dmxl1//Dmxl2//Ppt1//Slc11a1
positive_regulation_of_ATP_biosynthetic_process	0.00685437248156771	0.0355123288894215	2.16403229875441	Bcl2l1//Entpd5//Igf1//Insr//P2rx7//Stat3
positive_regulation_of_TOR_signaling	0.0102227149011714	0.0490207352347509	1.99043375097056	Akt3//Flcn//Gas6//Rictor//Rragc//Rragd//Smcr8
osteoclast_differentiation	0.0111237823954741	0.0520475125816834	1.95374751543618	Cd300lf//Ctnnb1//Fam20c//Fcer1g//Gab2//Ostm1//Tgfb1//Trf

Surprisingly, the pathway most impacted by low [Na⁺] turned out to be the lysosomal gene cluster with a strong 18-fold enrichment and 49 DEGs. This involvement of lysosomal response exceeded every other KEGG pathway change by low [Na⁺] compared with normal [Na⁺].

The gene ontology (GO) biological processes enrichment analysis results are congruent with the KEGG pathway analysis results. Findings revealed that low [Na⁺] exposure induces a significant overrepresentation of genes involved in cell cycle signaling, regulation of GTPase activity, osteoclast adhesion and motility, mitochondrial biogenesis and ATP production, and lysosomal functions including acid secretion, mTOR signaling and autophagy (Table 2).

3.1.2. Validation of RNA-sequencing data with microarray and qRT-PCR analyses

Murine whole genome microarray analysis results provided independent evaluation of differentially expressed protein coding transcripts in response to low [Na⁺] for 24 h in RAW264.7 cells. This analysis identified 2979 up-regulated and 2206 down-regulated transcripts in response to low [Na⁺]. Selected transcripts are listed in Table 3. Fold change was calculated as normalized gene expression in the low [Na⁺] samples divided by the normalized gene expression in normal [Na⁺] samples using a threshold of 2.0-fold change and p < 0.001. Positive values indicate up-regulation and negative values indicate down-regulation.

Table 3.

List of selected differentially expressed genes from mouse whole genome microarray analysis.

Gene name	Protein	Protein function	fold change	p-value
Up-regulated genes
Mmp14	matrix metalloproteinase 14	resorption	75.7	0.0056
MMP12	matrix metallopeptidase 12	invasion	65.1	0.0051
Gstt1	gluthation-S-transferase theta1	promotes osteoclast differentiation	45.9	1.12896815884778E-05
Traf1	TNF receptor associated factor 1	RANK signaling to NFκB	14.4	2.89036713230217E-07
Mmp9	matrix metalloproteinase 9	resorption	47	1 E–6
Apc2	adenomatous polyposis coli protein 2	activates Rho GTPase	15.9	0.5 E–3
Col2	collagen type 2 cleavage	resorption	15.8	0.0085
Ccl2	monocyte chemo attractant protein 1	osteoclastogenesis	15.2	0.005
Src	proto-oncogene tyrosine kinase	cell adhesion	8.7	0.011
Camk2d	calcium calmodulin dependent protein kinase 2 delta	regulates calcium and Na + influx	7.8	0.03
Atp6v0d2	vacuolar proton pump	acid production	5.3	7.84839173536556E-05
NFκB	nuclear factor kappa B	growth and differentiation	3.5	0.03
Ctsk	cathepsin K	resorption	2.8	8 E–09
Acp5	tartrate resistant acid phosphatase	resorption	2.3	1.1 E–09
Slc9A1	solute carrier transporter ATPase NCX1	lysosome function	4.2	0.05
Calcr	calcitonin receptor	resorption	1.6	0.02
Ank	ankyrin	Na ion transport, motility	14.7	2.5 E–7
Down-regulated genes
Aqp12	aquaporin 12	inhibits acid secretion	−14.5	0.1
Itga4	alpha 4 integrin	inhibits PI3K/Akt activity	−13.3	0.003
Mapk4	mitogen activated protein kinase 4	inhibits osteoclastogenesis	−11.4	0.06
Tspan10	tetraspanin 10	inhibits osteoclastogenesis	−7.1	0.03
Hpsa	heparanase	inhibits osteoclastogenesis	−2.9	0.0002

RNA was prepared from RAW264.7 osteoclastic cells exposed for 24h to media with low [Na⁺] (120 mmol/l) compared with expression from cells in normal [Na⁺] (140 mmol/l). Gene symbols, description, and function are listed along with the magnitude of mean expression changes. Variability within groups was less than 5%. Data confirmed results from RNA sequencing, implicating osteoclast specific genes, osteoclastogenesis, RANK pathway, MAP-kinase, lysosomal acid production, and cell adhesion.

Osteoclast specific gene expression changes were confirmed by separate qRT-PCR experiments. Among these genes, the expression of matrix metalloproteinase 9 (MMP9) increased by 47-fold; matrix metallopeptidase12 (MMP12) increased by 65-fold; matrix metalloproteinase 14 (MMP14) increased by 76-fold; cathepsin K (Ctsk) increased by 2.8-fold; tartrate resistant acid phosphatase (Acp5) increased by 2.3-fold; collagen type 2 cleavage (COL2) increased by 15.8-fold; and calcitonin receptor increased 1.6-fold, all compared to expression in cells cultured in normal [Na⁺] medium. Importantly, the upregulation of lysosomal vacuolar proton pump (Atp6v0d2) also increased significantly by 5.3-fold. The t-test p-values for all of these changes were below 0.01 for 3 samples each. Results from qRT-PCR tests performed after 48 h exposure to normal [Na⁺] or low [Na⁺] medium were not significantly different from results after 24 h exposure (data not shown).

Functional analysis of the microarray data confirmed findings obtained by analysis of RNA-sequence data. Here again, the KEGG pathway list was compiled from publicly available data and plotted against our experimental gene expression data. The −log p-values reflect the magnitude impact of gene expression changes associated with each osteoclast-specific processes. Table 4 lists the top cellular processes significantly affected by low [Na⁺] in the microarray experiments. Despite substantial differences caused by the better fidelity of RNA sequencing than the microarray analysis, there is a significant overlap in the pathways listed in Tables 2 and 4. Results unequivocally show the low [Na⁺] induced osteoclastogenic effects and increased osteoclast activities including acid secretion, motility, adhesion, and mitochondrial ATP production.

Table 4.

List of significantly affected pathways by low [Na⁺] based on data from whole genome microarray analysis.

PROCESS	Low [Na+]/Normal [Na+]
	% enrichment	−log p value
Sodium ion transport	100	129
Cell cycle regulation	68	50
Cell-matrix interactions	10.4	38
Osteoclast differentiation	16.8	31
Inhibition of Ca++ transport via voltage-gated calcium channel	28	31
NFκB pathway	39	27
Cell motility	29.8	24
Negative regulation of spermatogenesis	42	16
Skeletal development	29.3	14
ECM proteolysis	24	13
Positive regulation of cell proliferation	14.5	10
TGF-β Signaling	73	8
p53 signaling	68	8
Oxidative stress	70	7
Cell adhesion	30.8	7
Ossification	10	6
Osteoporosis	81	5
Angiogenesis	10	4
L-ascorbic acid transport	25	4

Results confirmed RNA-sequencing findings on top physiological processes selectively modified by for 24h exposure to low [Na⁺] (120 mmol/l) compared to normal [Na⁺] (140 mmol/l) in RAW264.7 murine osteoclastic cells. The −log p value above 1.5 is considered statistically significant. Processes are listed by −log p-value ranking.

3.2. Functional assays

3.2.1. Low [Na⁺] facilitated lysosome positioning and acidification

Acidic organelles were visualized by microscopy after loading with the acid sensitive dye DND-189 for 30 min in normal [Na⁺] medium or low [Na⁺] medium (Fig. 2A). Fluorescence intensity representing lysosome acidity is color coded according to the color bar on the left. Lysosome number, size, and acidity markedly increased in osteoclasts incubated in low [Na⁺] medium compared to lysosomes in osteoclasts incubated in normal [Na⁺] medium. Acidic lysosomes accumulated along the ruffled border in osteoclasts exposed to low [Na⁺] medium.

Fig. 2. Low [Na⁺] increased lysosomal acidification in RAW264.7 cell derived osteoclasts.

A. Microscopy depicts the increased acidity in vacuoles of osteoclasts exposed to low ECF [Na⁺] (LS, 120 mmol/l) that has been adjusted for equal osmolality compared to vacuoles in osteoclasts exposed to normal ECF [Na⁺] (NS, 140 mmol/l). Sealing zones of multinucleated osteoclasts are marked by arrows, decorated by bright green round acidic (pH < 4) lysosomes upon low ECF [Na⁺] exposure. Heat bar (left) indicates dye acidity in relative fluorescence units. Bars, 10 μm.

B. Graph depicts the time-dependent increase in acidity from RAW164.7 derived osteoclasts in response to low ECF [Na⁺] (LS, 120 mmol/l) exposure, compared to response to normal ECF [Na⁺] (NS, 140 mmol/l). Acidity was measured over time as intensity changes from cells loaded with the fluorescent dye LysoSensor Green DND-189 and the results are expressed as relative fluorescent units (FU) (mean ± SEM for 12 replicates). Green line shows data from LS and purple line shows data from NS exposed cells. Difference between the acidity in NS samples and acidity in LS control samples was statistically significant by 5 min (p < 0.005), the difference reached a maximum after 45 min (p < 0.0001) and remained steady after that for up to 2 h.

Assays done on a 96-well plates showed that the acidity increases after DND-189 addition (5 μl/well) were rapid and progressive (Fig. 2B). Serial measurements of DND-189 fluorescence intensity units (FU) demonstrated significant difference between acidity in normal [Na⁺] samples and acidity in low [Na⁺] control samples in 5 min (normal [Na⁺] 838 ± 138 FU, low [Na⁺] 1592 ± 182 FU; p < 0.005). The [Na⁺]-dependent difference in acidity reached a maximum after 45 min (normal [Na⁺] 3317 ± 158 FU, low [Na⁺] 5294 ± 338 FU; p < 0.0001) and remained steady after that up to 2 h.

Lysosomal acidity was increased by 19 ± 0.8% by incubation in low [Na⁺] for 24h compared to incubation in normal [Na⁺] (p < 0.001). RANKL exposure for 24h in normal [Na⁺] increased lysosomal acidity by 76 ± 2.8% compared to normal [Na⁺] without RANKL (p < 0.001). RANKL exposure for 48h in low [Na+] increased lysosomal acidity by 83 ± 3.1% compared to low [Na⁺] exposure without RANKL (p < 0.01). The difference between RANKL effect on lysosomal acidity in low [Na⁺] and RANKL effect in normal [Na⁺] was +14 ± 0.4% (p < 0.001).

3.2.2. Low [Na⁺] stimulated mitochondrial ATP production in differentiated osteoclasts

The upregulation of genes controlling mitochondrial biogenesis and ATP generation had physiological impacts, as measured by bioassays. Cytotoxicity indicator fluorescence units (FU) were similar in cells with low [Na⁺] media 22,788 ± 1,383 FU compared to cells in normal [Na⁺] media 21,064 ± 713 FU after 24h in culture (mean ± S.E; p = 0.29 NS from 6 replicates).

Bone resorbing activity in osteoclasts require high levels of ATP. ATP production is dependent on Src-kinase induced Cox activity; differential upregulation of Src mRNA by low [Na⁺] ATP generation was more intense in undifferentiated RAW264.7 osteoclastic cells cultured in low [Na⁺] media than in cells cultured in normal media, as shown in Fig. 3. ATP luminescence units were normalized for protein content of the lysate and showed higher ATP generation in cells after 24 h exposure to low [Na⁺] media (71.6 ± 5.6 FIU/μg protein) than after exposure to normal [Na⁺] media (33.4 ± 5.6 FIU/μg protein; data are mean ± S.E. from 6 replicates each; p < 0.001). Repeat experiments yielded similar results (not shown).

Fig. 3. Low ECF [Na⁺] stimulated mitochondrial ATP production in differentiated osteoclasts.

RAW264.7 monocytic cells were differentiated into mature osteoclasts for 11 days in media with cytokines, subcultured and then exposed to normal [Na⁺] (NS, 140 mmol/l) media and low [Na⁺] (LS, 120 mmol/l) media with cytokines for 24 h. Cells were then lysed and ATP content was measured using the Promega “ToxGlo” kit according to manufacturer’s protocol. Resulting ATP content is represented by luminescence units expressed as mean ± SEM of 6 replicates each. LS exposure significantly increased ATP levels (*p < 0.001 compared to NS).

3.2.3. Low [Na⁺] increased MMP9 expression and facilitated osteoclast migration

Fig. 4A depicts Western blot analysis using Abcam MMP-9 antibodies and compares expression changes induced by low [Na⁺] compared with normal [Na⁺], as well as changes with RANKL addition and without RANKL addition, each. The upper 92 kDa protein bands represent proMMP-9 and the lower 82 kDa bands represent active MMP-9. Relative expression was enumerated with densitometry and are included in parenthesis as follows. Incubation with low [Na⁺] for 24h (Lane 3, RE 28) and 48h (Lane 7, RE 38) did not substantially increase proMMP-9 expression compared to normal [Na⁺] for 24h (Lane 1, RE 23) and 48h (Lane 5, RE 33). RANKL addition markedly increased proMMP-9 expression after 24h in normal [Na⁺] (Lane 2, RE 98) and even more after 48h (Lane 6, RE 152). RANKL induced increase of proMMP-9 expression in low [Na⁺] media was augmented after 24h (Lane 4, RE 136) and 48h (Lane 8, RE 172) compared to the expression in normal [Na⁺] after 24h (Lane 2, RE 98) and 48h (Lane 6 RE 152). These results suggest a synergistic effect between low [Na⁺] and RANKL to induce proMMP-9 gene expression.

Fig. 4. Low ECF [Na⁺] increased MMP-9 expression and facilitated osteoclast cell migration.

A. MMP-9 expression was independently increased by low [Na⁺] and RANKL treatment. Representative MMP-9 Western blot using the Abcam antibodies. Lysates from 24h incubation: Lane 1: normal [Na⁺], Lane 2: normal [Na⁺] with RANKL, Lane 3: low [Na⁺], and Lane 4: low [Na⁺] with RANKL. Lysates from 48h incubation: Lane 5: normal [Na⁺], Lane 6: normal [Na⁺] with RANKL, Lane 7: low [Na⁺], and Lane 8: low [Na⁺] with RANKL. The upper bands represent proMMP-9 expression and the lower bands represent the active MMP-9 after cleavage by the lysosomal enzyme cathepsin K.

B. RAW264.7 monocytic cells were differentiated into mature osteoclasts by culturing in media with cytokines for 11 days and then transferred into Boyden chambers filled with either normal ECF [Na⁺] (NS, 140 mmol/l) or low ECF [Na⁺] (LS, 120 mmol/l) using the Cultrex 96-well Kit (Trevigen Inc.). Cells were allowed to migrate through the membrane for 24 h, then migrated cells were counted. Data are mean ± S.E.M. from 6 replicates each (*p < 0.01 compared to NS).

The activated MMP-9 expression (see lower bands in Fig. 4A) was increased by low [Na⁺] for 24h (Lane 3, RE 13) compared to the expression in normal [Na⁺] (Lane 1, RE 7). This increase was slightly higher after 48 h in low [Na⁺] (Lane 7, RE 16) and normal [Na⁺] (Lane 5, RE 16). RANKL addition increased conversion of inactive to active MMP-9. Exposure to RANKL induced MMP-9 in low [Na⁺] for 24h (Lane 4, RE 84) and 48h (Lane 8 RE 91) more than RANKL in normal [Na⁺] for 24h (Lane 2, RE 33) and for 48h (Lane 6, RE 63). Results using antibodies from R&D Systems matched the results obtained using Abcam antibodies (data nor shown).

Osteoclast migration is strongly supported by proteolytic activity, and the migration phase allows osteoclast to select a new resorption location. This migration is also required for invasion of cancer cells. Migration through a membrane is a well-established model of osteoclast motility (Golden et al., 2015; Xuan et al., 2017). RAW264.7 cells were exposed to culture media with normal [Na⁺] = 140 mmol/l or low [Na⁺] = 120 mmol/l) with cytokines for 24 h in Boyden chambers over the 24 h incubation. Migrated cells through the membrane into the lower chamber were counted. In normal media, 28 ± 6% of cells migrated across the membranes, whereas in low [Na⁺] medium 47 ± 2% cells migrated (p < 0.01), as shown on Fig. 4B. Results were similar in 4 additional repeated experiments.

4. Discussion

Results highlight extensive low [Na⁺] induced changes in osteoclast gene expression (Fig. 1 and Table 1). Exposure to low [Na⁺] induced differential gene expression that accounts for accelerated osteoclast formation, osteoclast differentiation and increased osteoclast activities. Therefore, our findings conclusively identify osteoclasts as a direct target of low ECF [Na⁺].

Functional enrichment analysis of differentially expressed genes (Table 2) demonstrated that low [Na⁺] primarily impacts lysosomal signaling. Recent research elucidated the pivotal roles of lysosomal signaling in the regulation of osteoclast functions (Lacombe et al., 2013; Wang et al., 2017; Hu et al., 2016; Lawrence and Zoncu, 2019), showing that lysosomes serve as a cellular center for signaling and metabolism through interactions with the mTOR complex to participate in nutrient and growth factor sensing. Lysosomes degrade intracellular and extracellular macromolecules using proteases, lipases, nucleases and other hydrolytic enzymes in osteoclasts (Lacombe et al., 2013; Erkhembaatar et al., 2017; Hu et al., 2016; Teitelbaum, 2011). This catabolism requires acidic pH, which is established by an ATP-driven proton pump, the vacuolar ATPase in cooperation with ion channels. Our results show that low [Na⁺] selectively upregulates genes involved in this degradative machinery, including an upregulation of v-ATPase expression. The functional significance of lysosomal gene expression changes is supported by our bioassays, demonstrating that low [Na⁺] time-dependently increased lysosomal acidity and triggered migration of acidic lysosomes to the secretory membrane.

GO term physiological processes that were highly affected by low [Na⁺] included mitochondrial ATP/energy production (Fig. 3) and increased osteoclast motility (Fig. 4). This is consistent with the results of our gene expression analysis demonstrating that low [Na⁺] positively regulates mitochondrial ATP biogenesis (6 genes), leading to a burst in energy metabolism. This induction of ATP production is signified by the bioassay results. Combined, these data therefore support the hypothesis that hyponatremia activates osteoclasts to liberalize sodium from bone matrix stores via bone resorption (Barsony et al., 2012).

Low [Na⁺] induced gene expression changes strongly correlated with osteoclast functional changes, including increased lysosomal acidity (Fig. 2), stimulation of mitochondrial energy production (Fig. 3), marked increase in MMP-9 expression (Fig. 4A), and osteoclast motility (Fig. 4B). Previously, we also found that low [Na⁺] exposure increased osteoclast proliferation and differentiation, all consistent with functional analysis of low [Na⁺] induced differentially expressed genes in our mRNA sequencing experiment.

Pathway analysis of differentially expressed genes confirmed and extended our prior findings. Activation of these pathways is known to increase in osteoclast formation, differentiation and osteoclast specific functions (Teitelbaum, 2007). Our bioassays supported new aspects of osteoclast activation by low [Na⁺], including increased lysosomal acid production and polarized localization (Fig. 2A and B). Our prior studies showed low [Na⁺] dependent p53 protein overexpression, NF-kB p65 phosphorylation, and S6 ribosomal protein phosphorylation. Analysis of sequencing data now show that significantly affected pathways include osteoclast nutrient sensing PI3K/Akt/mTOR pathway, MAPK and NF-kappa B pathways, ascorbate pathway and mitochondrial pathway genes, ion channels, and G-protein coupled receptors, as shown in Table 3. We also found that the low [Na⁺] induced lysosomal signaling include the dysregulation of autophagy.

Treatment of differentiated cells with RANKL in low [Na⁺] increased lysosomal acidity and MMP-9 expression synergistically compared to the increases by low [Na⁺] or RANKL alone. This result supports our prior findings of independent increases in osteoclast differentiation and bone resorption in RAW264.7 cells and primary osteoclasts exposed to low ECF [Na⁺] (Barsony et al., 2011), and is now explained by the low [Na⁺] induced mitochondrial energy production.

These signaling changes likely contribute to low [Na⁺] induced osteoporosis (Florencio-Silva et al., 2017) and allow predictions for potential molecular targets for treatment. The involvement of mTOR signaling in low [Na⁺] effects in osteoclasts predicts that mTOR inhibitors, such as everolimus (Hadji et al., 2013; Hadji et al., 2013; Cejka et al., 2010), may be potential agents for the prevention or treatment of hyponatremia-induced osteoporosis.

Interestingly, pathway analysis revealed that low [Na⁺] induces overexpression of the Ras-related GTPases and associated proteins, which is consistent with low [Na⁺] induced G-protein coupled receptor signaling. Future studies will likely identify a low ECF sodium sensing and responding plasma membrane receptor as a fundamental component of low [Na⁺] signaling in osteoclasts, and may suggest additional potential antagonists (Stockert and Devi, 2015; Tang et al., 2012).

This report marks a significant step toward understanding low [Na⁺] cellular signaling and confirms osteoclasts as target cells for low [Na⁺] effects. Nevertheless, several important aspects of the low [Na⁺] induced signaling remain to be elucidated by studies that go beyond this initial study. Here we utilized cultured murine RAW264.7 osteoclastic cells with exposure to low [Na⁺] for 24h. The RAW264.7 cell line is derived from the monocyte/macrophage lineage and is capable of differentiating into osteoclasts and resorbing bone. Future gene expression analysis expected to include primary murine osteoclasts and human osteoclasts. Finding direct effects of low [Na⁺] on human osteoclasts would support the hypothesis that our prior rodent studies are relevant to human physiology, as suggested by epidemiological studies in human populations (Usala et al., 2015). In addition, we suggest the possibility that the low [Na⁺] impact on lysosomal signaling may be a general characteristic across cells and species, which would indicate a pivotal role of lysosomal signaling in low [Na⁺] induced effects on other cells and organs.

Our studies on transcriptomic changes may be extended to include initial changes of gene expression and protein interactions in minutes and hours. Previous efforts to characterize rapid low [Na⁺] responses were confounded by concomitant osmolality changes, which resulted in cell swelling and detrimental effects on cell function. Our discovery that we can compensate for the osmolality changes by adding mannitol without needing to use our previously employed method of slowly adapting osteoclastic cells to low extracellular [Na⁺] via stepwise reduction in the medium osmolality over 10–14 days (Barsony et al., 2011) has now allowed us to characterize a broad range of osteoclast responses to low [Na⁺] much more specifically and quickly. We used mannitol to compensate for low osmolality in the low sodium medium. Several reports indicated that mannitol itself could influence osteoclast formation and gene expression in RAW264.7 cells. Considering that this effect of mannitol was inhibitory to osteoclast growth and differentiation (Tanaka et al., 2017), and suppressed osteoclastogenesis by inhibiting expression of the osteoclast specific gene MMP-9 (Bai et al., 2005), it is unlikely, that the use of mannitol to correct osmolality in the low sodium medium would prevent us from exploring rapid effects on osteoclast signaling. The specificity of low [Na⁺] induced changes in our experiments were strictly controlled by matching media content in all samples regarding ionic content (except sodium), osmolality, nutrients, cytokines, vitamins, antioxidants.

Our findings have far-reaching impact for future studies on hyponatremia-induced osteoporosis. First, the finding of murine cell responses allows utilization of genetically modified mouse models to study key elements in the pathways responsible for osteoclast formation and activation in hyponatremia-induced osteoporosis, including questions on prevention, reversibility, and treatment. Moreover, fundamental knowledge is lacking regarding cellular adaptation to low sodium in the environment. Our future mechanistic studies at the cellular level will provide insight into sodium conservation at the cellular level. These studies are expected to include proteomic and pharmacological approaches to complement our understanding of evolutionary and pathological aspects of sodium and water homeostasis.

Abbreviations

AVP

arginine vasopressin

ATCC

American Type Culture Collection

ATP

adenosine triphosphate

Akt

RAC-alpha serine/threonine-protein kinase (Protein Kinase B)

BMM

bone marrow macrophages

FU

fluorescence unit

low [Na⁺]

low extracellular fluid sodium ion concentration

M-CSF

macrophage colony stimulating factor

αMEM

Minimal Essential Media Alpha

mTOR

mammalian target of rapamycin

PBS

phosphate buffered saline

PI3K

phosphoinositide-3-kinase

PTH

parathyroid hormone

RANKL

receptor activator of nuclear factor kappa-B ligand

RT-PCR

reverse transcription polymerase chain reaction

SIADH

syndrome of inappropriate antidiuretic hormone secretion

TNF

tumor necrosis factor

TRAP

tartrate resistant acid phosphatase

V-ATPase

vacuolar-type proton pump adenosine triphosphatase