Effects of acid on bone

Abstract

The homeostatic regulation of a stable systemic pH is of critical importance for mammalian survival. During metabolic acidosis (a reduction in systemic pH caused by a primary decrease in serum bicarbonate concentration), as seen in clinical disorders such as the later stages of chronic kidney disease, renal tubular acidosis, or chronic diarrhea, bone buffers the accumulated acid; however, this homeostatic function of the skeleton occurs at the expense of the bone mineral content and leads to decreased bone quality. During acute metabolic acidosis, there is initial physiochemical bone mineral dissolution, releasing carbonate and phosphate proton buffers into the extracellular fluid. In addition, there is net proton influx into the mineral with release of bone sodium and potassium. During chronic metabolic acidosis, there is also inhibition of osteoblast activity, resulting in reduced bone formation, and an increase in osteoclast activity, resulting in increased bone resorption and release of calcium and anionic proton buffers. These physicochemical and cell-mediated bone responses to metabolic acidosis, in addition to an acidosis-induced increased urine calcium excretion, without a corresponding increase in intestinal calcium absorption, induce a net loss of body calcium that is almost certainly derived from the mineral stores of bone.

Metabolic acids are derived from the metabolism of dietary proteins, phospholipids, and nucleic acids and result in a daily addition of ≈0.7 to 1.0 mEq/kg per day to the systemic circulation.^1,2 These acids are buffered, before their excretion, to prevent a decrease in extracellular pH.³ Both intracellular buffers, such as anionic proteins and phosphates, especially in muscle, and extracellular buffers, principally bicarbonate in addition to the phosphates and carbonates in bone, help to mitigate a decrease in extracellular pH. In patients with normal kidney function, these metabolic acids are soon quantitatively excreted through enhanced ammonia and titratable acid excretion. The level of serum bicarbonate, which is generally used clinically to measure systemic acid base status, is maintained within the normal range. However, with a decline in kidney function, the daily addition of metabolic acids begins to exceed the capacity of the kidney to quantitatively excrete these metabolic acids and they begin to accumulate. Initially, further acid (proton) buffering prevents a decrease in serum bicarbonate below the lower limit of the normal range (generally considered to be 22–29 mEq/L). Acid retention with the concentration of serum bicarbonate in the normal range is termed eubicarbonatemic (or normobicarbonatemic) metabolic acidosis.^1,4 However, although within the normal range, the level of serum bicarbonate may be lower than the previous baseline for that individual. Urine excretion of citrate (an abundant organic base) and ammonia decreases with progressive chronic kidney disease (CKD).^5–7 This eubicarbonatemic metabolic acidosis contributes to progression of CKD.^8–10 With further loss of kidney function, as the proton buffering capacity of the systemic buffers is overwhelmed, the serum concentration of bicarbonate decreases below the lower limit of the normal range and criteria for the clinical definition of metabolic acidosis (a reduction in systemic pH caused by a primary decrease in serum bicarbonate concentration) are satisfied.³ In this review, we will elaborate how metabolic acidosis affects the bone mineral, especially carbonates and phosphates, and osteoblastic (responsible for bone formation) and osteoclastic (responsible for bone resorption) function as the bone buffers the additional protons.

In vivo studies

There is substantial in vivo evidence that metabolic acidosis adversely affects bone.^11–16 Approximately 60% of the increase in protons is acutely buffered by bone^{12,13,15–23} and soft tissues.^16,24–27 Animal models of acute metabolic acidosis demonstrate that sodium, potassium, carbonate, and calcium are lost from bone during acid exposure, suggesting the exchange of sodium and/or potassium for protons on the bone surface and consumption of bone mineral carbonate in buffering the excess protons.^{20,22,28–41} In acidemic patients with CKD, bone carbonate is decreased,^42–44 which may represent dissolution of carbonate stores and/or replacement by phosphate as protons are incorporated into the mineral^20,22,35 and the rate of fracture is increased.^45–48 In most in vivo studies, chronic metabolic acidosis is associated with a concomitant decrease in bone mineral content,^49,50 which appears to reflect acid-mediated dissolution and/or resorption of bone mineral.^{12,17,41,49,51–54} The net negative calcium balance is due to increased urine calcium excretion, without a corresponding increase in intestinal calcium absorption,^53,55,56 resulting in a net loss of body calcium.^16,50,57 Bone is almost certainly the source of this additional urinary calcium, as ~98% of body calcium is contained within the skeleton.⁵⁸ The administration of oral base as either bicarbonate or citrate has been shown to correct acidosis-induced bone mineral loss.^59–63

The metabolism of dietary protein generates metabolic acids that may be increased by consumption of the common high endogenous acid-producing diet of North Americans, coupled to the age-related decline in kidney function and ability to excrete acid.^14,64,65 This has led to the hypothesis that acid retention and its resorptive effect on bone play a role in causing osteoporosis.^{2,14,15,49,66–68} This hypothesis is supported by the observation that administration of base decreases the negative calcium balance induced by a high protein diet.^69–71 Postmenopausal women were fed a constant diet, and their endogenous acid production was neutralized with potassium bicarbonate.⁷⁰ In the test subjects, potassium bicarbonate reduced urinary calcium excretion, making the calcium balance less negative. Bicarbonate treatment reduced urinary excretion of hydroxyproline, a marker of bone breakdown, and increased excretion of serum osteocalcin, a marker of osteoblastic bone formation. These results have been confirmed in other studies in postmenopausal women and in elderly men and women.^{62,63,72–74} In a randomized, prospective, controlled, double-blind trial, 161 postmenopausal women with low bone mass received 30 mEq of either oral potassium citrate or potassium chloride as a control. The women who received potassium citrate had a significant intergroup increase in bone mineral density (±SE) of 1.87% ± 0.50% at L2 through L4 (P < 0.001), 1.39% ± 0.48% (P < 0.001) at femoral neck, and 1.98% ± 0.51% (P < 0.001) at total hip.⁶² In another randomized, double-blind, placebo-controlled trial, 60 mEq of potassium citrate given to 201 elderly healthy men and women for 24 months significantly increased areal bone mineral density at the lumbar spine (L2–L4) by 1.7% (95% confidence interval, 1.0%–2.3%) net of placebo and in the femoral neck by 1.6% (95% confidence interval, 1.1%–2.2%) net of placebo. In this study, potassium citrate significantly decreased the bone fracture prediction score by FRAX (fracture risk assessment tool that predicts the percentage probability of any fracture in 10 years) analysis in both men and women.⁶³

Metabolic acidosis also alters new bone formation. The independent effect of acidosis to suppress bone formation in vivo has been demonstrated in children with distal renal tubular acidosis.^75,76 Alkali therapy helped these children attain and maintain normal height. The administration of alkali to correct the acidosis improved bone histology and mineral density in adults with distal renal tubular acidosis.^77,78 In a radiographic study, most patients with proximal renal tubular acidosis had rickets or osteopenia.⁷⁹ In vitro studies support these in vivo observations and provide insight into the mechanisms by which bone buffers the excess protons during acid retention and will be reviewed below.

Acute metabolic acidosis

Calcium release.

In vitro organ culture and cell-based assays have been used extensively to study calcium release and proton buffering in response to models of acute metabolic acidosis. Neonatal mouse calvariae (frontal and parietal bones of the skull) contain functioning osteoblasts and osteoclasts, respond to hormones, and synthesize DNA and protein like human bone.^23,80,81 The calvariae can be cultured in the physiological carbon dioxide–bicarbonate buffered medium, allowing the pH to be independently altered within the physiological range by changing the bicarbonate concentration (simulating metabolic acidosis or alkalosis) or the partial pressure of carbon dioxide (simulating respiratory acidosis or alkalosis). In this model, cultured calvariae exhibit proton-dependent net calcium efflux during both acute (≤3 hours) and chronic (≥24–99 hours) incubations in medium modeling metabolic acidosis (Figure 1).^23,82–84

Figure 1 |. Response of bone to metabolic acidosis.

Acidosis has both short-term (<24 hours; blue) and long-term (>24 hours; red) effects on bone. Short-term responses are primarily physicochemical exchanges of protons (H⁺) with sodium (Na⁺) and potassium (K⁺) as well as Ca release; long-term responses are cell mediated and result, in addition, to the release of H⁺ buffers carbonates (CO₃²⁻) and phosphates (PO₄^3–). Osteoblast activity is regulated, leading to decreased bone formation and increased signaling that activates osteoclastic bone resorption. Collagen, alkaline phosphatase (AP), osteopontin (OP), and matrix Gla protein (MGP) are decreased, whereas prostaglandin E₂ (PGE₂), receptor activator of nuclear factor-κB ligand (RANKL), and fibroblast growth factor 23 (FGF23) are stimulated. Osteoclasts also respond directly to H⁺ in addition to the osteoblast-mediated activation through the receptor activator of nuclear factor-κB (RANK), leading to an increase in β-glucuronidase (β-Glu).

In acute (3-hour) experiments in neonatal mouse calvariae, there was a consistent slope between the medium pH and the calcium flux, regardless of treatments that stimulated or suppressed bone cell activity without directly affecting the bone mineral.⁸² These findings demonstrated that acute acid-mediated calcium release from bone is primarily due to physicochemical, and not cell-mediated, mechanisms. Synthetic, carbonated apatite disks, a cell-free model of bone mineral,^85–87 immersed in a physiologically acid medium, demonstrated similar net calcium efflux to that observed in calvariae over a 3-hour period of incubation,⁸⁸ supporting a physicochemical mechanism for the acute response of bone to protons.

To determine the type of bone mineral that would be affected by physicochemical forces, phosphate or carbonate concentrations were altered at either neutral or physiologically acid medium pH in calvarial cultures and the net calcium flux was measured.²² With respect to calcium and carbonate, but not calcium and phosphate, there was bone formation in a medium supersaturated with calcium carbonate, no change in the bone mineral when cultured in a saturated medium, and bone dissolution into an undersaturated medium. This indicated that bone carbonate appears to be selectively solubilized during an acute reduction in pH, leading to a release of calcium. Indeed, there was a progressive loss of total bone carbonate when calvariae were cultured in acidic medium.²⁰ At a constant pH, whether physiologically neutral or acid, net calcium efflux from bone is dependent on the medium bicarbonate concentration; the lower the bicarbonate concentration, the greater the calcium efflux.⁸⁹ Bone carbonate appears to be in the form of carbonated apatite.^90,91

Hydrogen ion buffering.

Using in vitro neonatal mouse calvariae, incubation in culture medium acidified by a decrease in bicarbonate concentration resulted in a net influx of hydrogen ions into the bone, increasing the pH of the culture medium and indicating that hydrogen ions were buffered by bone.^21–23,27 In addition to calcium release from bone in response to metabolic acidosis, extensive in vitro evidence indicates acid buffering by bone is due to 2 key mechanisms: exchange of sodium and/or potassium for protons and consumption of carbonate and phosphate buffers released from the bone.^{20–23,27–29,33,34,83,89,92–94}

Bone is a reservoir for sodium and potassium, and its surface has fixed negative sites that normally complex with sodium, potassium, or hydrogen ions, with the sodium and potassium exchanging freely with the surrounding fluid.²⁷ This surface distribution can be altered in the presence of additional hydrogen ions to lower the bicarbonate concentration and pH as a model of metabolic acidosis.^27,83 A high-resolution scanning ion microprobe was used to determine that incubation in acidic medium caused a loss of surface sodium and potassium relative to calcium in conjunction with hydrogen ion buffering, suggesting that sodium and potassium exchanged with hydrogen ions on the bone surface, resulting in a decrease in medium acidity.^28,29,33,34 This model is consistent with animal studies of acute acidosis using the stable isotope ⁴¹K to measure the ⁴¹K-to-calcium ratio, which demonstrated that acidosis induces a decrease in this ratio, suggesting a loss of the stable potassium isotope from the bone mineral as a consequence of acid exposure.³⁴

Carbonate and phosphate buffers are released from bone in response to models of metabolic acidosis. Bone contains 80% of the total carbon dioxide content of the body, and acute acidosis decreases total bone carbon dioxide.^35,95 Bone also contains ≈90% of total body phosphate, largely in the form of hydroxyapatite and derivative forms of apatite.^45,96 During acidosis, protonation of the phosphate in apatite consumes protons and restores pH to the normal range.^15,16,27,50 In addition, acidosis releases bone calcium and carbonate, leading to a progressive loss of bone carbonate over time.^20,22 On the bone surface, there is a 4-fold depletion of carbonate in relation to phosphate, and, in cross-section, there is a 7-fold depletion of phosphate in relation to carbonate. These results indicate that acute hydrogen ion buffering by bone involves preferential dissolution of surface carbonate and of cross-sectional phosphate.⁹² Thus, during acute metabolic acidosis, sodium and potassium are exchanged for protons on the bone surface, and bone carbonate and phosphate are lost from the bone mineral, all to help neutralize the additional protons and help maintain a physiological normal pH.^16,27

Relationship between calcium release and hydrogen ion buffering.

As described above, during acute metabolic acidosis, a reduction in pH causes both bone calcium release and proton buffering. With simple mineral dissolution, there should be a 1:1 ratio of protons buffered to calcium released in the case of calcium carbonate, 5:3 for apatite, and 1:1 for brushite.⁹⁷ However, with cultured calvariae, the ratio was found to be 16 to 21:1, indicating that proton buffering could not simply be due to mineral dissolution.²³ This indicates that calcium release is only one component of proton buffering by bone and is substantiated by the ion microprobe studies that show sodium and potassium exchange for protons^{28,29,33,34,93} and loss of bone phosphate and bicarbonate with acidosis.^94,98

Chronic metabolic acidosis

Alteration in bone cell activity.

The chronic response (≥24 hours) to metabolic acidosis is dependent on specific bone cell responses; net calcium efflux from bone is mediated by inhibition of osteoblast activity, leading to decreased bone formation, and regulation of signaling, leading to increased osteoclast activity and enhanced bone resorption (Figure 2).^{27,83,99–104} Increased resorption leads to release of proton buffers from bone, and decreased formation leads to less incorporation of proton buffers into the mineral. Conversely, an increase in medium bicarbonate concentration (modeling metabolic alkalosis) decreases calcium efflux from bone through increased osteoblastic bone formation and decreased osteoclastic bone resorption.^21,105,106

Figure 2 |. Cellular response to metabolic acidosis.

Protons (H⁺) interact with a specific receptor, ovarian cancer G-protein coupled receptor 1 (OGR1), in both the osteoblast and the osteoclast. This receptor is coupled to phospholipase C (PLC), and its activation leads to inositol trisphosphate (IP₃)–mediated release of intracellular calcium (Ca²⁺), which can alter specific gene transcription in the osteoblast. The transcription factor early growth response 1 (Egr1), matrix Gla protein (MGP), and osteopontin (OP) are decreased in response to metabolic acidosis, whereas cyclooxygenase 2 (COX2) is increased, catalyzing increased prostaglandin E₂ (PGE₂) production from arachidonic acid (AA). PGE₂ then stimulates production of receptor activator of nuclear factor-κB ligand (RANKL) as well as fibroblast growth factor 23 (FGF23) by paracrine activation of the EP4 prostaglandin receptor. RANKL is secreted and binds to its receptor, receptor activator of nuclear factor-κB (RANK), on the osteoclast, leading to increased bone resorption. Release of the lysosomal enzyme, β-glucuronidase is indicative of osteoclast activation. The specific intracellular signaling initiated by OGR1 in the osteoclast is less well understood, but both pathways are necessary for the full response to metabolic acidosis.

Decreased osteoblast activity.

After 24 to 48 hours in culture, neonatal mouse calvariae incubated in medium with a physiological reduction in bicarbonate and pH, as a model of metabolic acidosis, demonstrate a marked decrease in osteoblast secretion of alkaline phosphatase, reduced synthesis of collagen, and reduced expression of noncollagenous proteins, such as osteopontin and osteocalcin, which in aggregate lead to reduced bone mineralization.^{27,99,107–109} Bone responds to chronic metabolic acidosis by reducing the content and production of new osteoblasts and by generating osteoblastic signaling molecules that activate osteoclasts and bone resorption.^27,103 Osteoblast collagen synthesis and alkaline phosphatase activity were both decreased compared with controls in calvarial cultures.^99,107 Egr-1,a key transcription factor involved in the maturation of fibroblasts, connective tissue, and bone cells,¹¹⁰ is inhibited by chronic metabolic acidosis.¹⁰⁹ In contrast, expression of c-fos, c-jun, junB, and junD RNA in osteoblasts was not affected by a similar decrement in medium pH.¹⁰⁷ In a model using primary cells isolated from neonatal mouse calvariae, consisting primarily of osteoblasts, incubations up to 23 days in duration in a physiologically acidic medium reduced collagen synthesis, formation of apatite nodules, and influx of calcium into the apatite nodules, compared with incubation in neutral medium.¹¹¹ In addition, in incubations up to 44 days in duration, bone matrix protein gene expression (e.g., matrix Gla protein and osteopontin) was also reduced on exposure to physiologically acidic medium relative to expression in neutral medium.¹⁰⁹ Similar inhibition of mineralization by acidosis in vitro has also been demonstrated in cultured rat calvarial osteoblasts¹¹² and human osteoblasts.¹¹³ Studies of the effects of acidosis on bone in intact animals and humans are needed to confirm that these changes observed in vitro also occur in vivo.

The inhibition of osteoblast activity by metabolic acidosis is mediated by the proton-sensing receptor, ovarian cancer G-protein coupled receptor 1, OGR1 (also called GPR68).^103,114 Ludwig et al. first reported that OGR1 is expressed in mouse osteoblasts, and a reduction in pH led to an accumulation of phosphoinositide metabolites.¹¹⁵ OGR1 has also been found in osteoclasts.^116–119 One of the first intracellular responses in the osteoblast to metabolic acidosis is an increase in inositol trisphosphate–mediated intracellular calcium that was inhibited by CuCl₂, a known inhibitor of OGR1.¹¹⁴ Pharmacologic blockade of inositol trisphosphate–mediated intracellular calcium was shown to inhibit acid-induced bone resorption, suggesting that this initial signaling response is critical for the response of the osteoblast to metabolic acidosis and subsequent stimulation of osteoclastic bone resorption.¹²⁰ In rapidly growing mice that generate significant metabolic acids, global deletion of OGR1 in male mice was found to result in increased bone density and increased osteoblast gene expression with a greater increase in bone formation than bone resorption.^103,104 In this report, female mice were not studied. These results indicate that OGR1 is important in the initial response of bone to metabolic acidosis.

In addition to inhibiting intracellular responses in the osteoblast that lead to an inhibition of osteoblastic bone formation, metabolic acidosis has been shown to increase levels of prostaglandins in a variety of model systems.²⁷ Prostaglandins, particularly prostaglandin E₂ (PGE₂), are potent multifunctional regulators of bone formation and resorption.^121–123 PGE₂ has been shown to promote new bone formation in some animal models and in a cell culture of isolated osteoblasts.¹²¹ In a neonatal mouse calvarial cell culture model, stimulation of PGE₂ is an important intermediate step in the response of osteoblasts to proton-mediated synthesis and secretion of receptor activator of nuclear factor-κB ligand (RANKL).^27,83,124 PGE₂ stimulates bone resorption in organ culture and mediates bone resorption of mouse calvariae in response to a variety of cytokines and growth factors.^27,123 In addition, the inhibition of PGE₂ production by indomethacin limits acid-induced bone calcium release and acid stimulation of RANKL in the osteoblast.^{100,123,125,126} Cortisol also inhibits acid-induced bone resorption through a decrease in osteoblastic PGE₂ production.¹²⁷ Metabolic acidosis stimulates the expression of several genes in osteoblasts, including cyclooxygenase 2, the rate-limiting step in prostaglandin production.^100,128 Pharmacologic blockade of cyclooxygenase 2 synthesis by the specific inhibitor NS-398¹²⁹ prevents the downstream increase in RANKL and acid-induced bone resorption.¹²⁸ In addition, mice with a genetic deficiency in cyclooxygenase 2 demonstrate a reduced response to acid-induced bone resorption.¹²⁸ These experimental results indicate that acid-induced, cell-mediated calcium efflux from bone is regulated downstream of OGR1, at least in part, by an increase in endogenous PGE₂ production, which, in turn, activates RANKL expression in osteoblasts to stimulate osteoclastic bone resorption and promotes release of bone calcium and anionic buffers, the latter of which help to neutralize the retained acid.

Increased osteoclastic activity.

The release of osteoclastic β-glucuronidase, a lysosomal enzyme whose secretion correlates with osteoclast-mediated bone resorption, is increased during culture of neonatal mouse calvariae in acid medium compared with neutral medium.^99,130 Osteoclast activation is critical for metabolic acidosis to induce bone resorption. The molecular mechanisms for this activation appear to require initial osteoblast signaling; however, recent findings suggest that there is also a direct effect of metabolic acidosis on the osteoclast.¹³¹

The important role for the osteoclastic proton receptor OGR1 in the induction of osteoclastic bone resorption has been examined with the use of osteoclast-specific OGR1 knockout (OC-cKO) female mice.¹³¹ Micro–computed tomography demonstrated increased density in tibiae and femurs but not vertebrae of these OC-cKO female mice, consistent with previous work using a global deletion of OGR1 in male mice.¹⁰³ In experiments using the OC-cKO mice where there is normal content of OGR1 in osteoblasts, isolated, differentiated osteoclasts derived from the bone marrow of femurs from OC-cKO mice demonstrated a decreased number of osteoclasts, pit formation, and osteoclast-specific gene expression compared with wild-type osteoclasts.¹³¹ The global OGR1 knockout mice showed histologic evidence of increased osteoclast activity, although subsequent characterization of isolated osteoclasts from global knockout bone demonstrated decreased osteoclast gene expression comparable to that observed in OC-cKO osteoclasts. In response to metabolic acidosis, osteoclasts from OC-cKO mice had decreased nuclear translocation of NFATc1, a transcriptional regulator of differentiation, and no increase in osteoclast size or number, as observed in osteo-clasts from wild-type mice. Thus, loss of osteoclast OGR1 decreased both basal and metabolic acidosis-induced osteo-clast activity, indicating that a direct effect on osteoclast OGR1 is important in mediating metabolic acidosis-induced bone resorption in addition to the responses of OGR1 in the osteoblast. These results are generally consistent with a direct role for OGR1 in the osteoclast, which has also been suggested by several other groups,^{116–119,132,133} although many of the critical experiments did not use a physiological model of metabolic acidosis; the pH was below the physio-logical range and/or buffers other than physiological bicar-bonate/carbon dioxide were used. In our study, only OC-cKO female mice, and not male mice, exhibited a significant bone phenotype.¹³¹ This gender difference is consistent with other studies that have shown a differential effect on bone mass based on sex in a large number of individual gene knockouts.^134,135 Future studies will be needed to address this difference in bone phenotype between female and male mice.

Effects of metabolic acidosis on bone in patients with CKD

Osteoblast inhibition and osteoclast activation in metabolic acidosis are 2 bone responses that, together with selective gene activation, regulate calcium, carbonate, and phosphate release to neutralize the additional acid load (Figure 3). In vivo, there are potential hormonal responses that can impact the effects of metabolic acidosis on bone, especially in the presence of CKD, including autocrine and paracrine effects of parathyroid hormone (PTH), fibroblast growth factor 23 (FGF23), 1,25-dihydroxyvitamin D₃, and PGE₂.^{15,83,136–138} The resorptive state of the bone is also likely to influence the magnitude of acid buffering by bone.¹³⁹ The simultaneous presence of CKD, metabolic acidosis, and osteoporosis often reinforces the detrimental effects of these hormones, increasing bone loss.

Figure 3 |. Contribution of metabolic acidosis to chronic kidney disease (CKD)–mineral and bone disorder.

As kidney function declines, renal acid excretion becomes quantitatively less than endogenous acid production, leading to chronic metabolic acidosis. Metabolic acidosis is among the factors that lead to a reduction in levels of serum 1,25-dihydroxyvitamin D₃ (calcitriol) and an increase in levels of parathyroid hormone (PTH) and fibroblast growth factor 23 (FGF23). Metabolic acidosis also accelerates the progression of CKD. The osseous response to metabolic acidosis results in the release of calcium (Ca²⁺) and the proton (H⁺) buffers carbonates (CO₃²⁻) and phosphates (PO₄³⁻), which help restore the systemic pH toward normal. There is also exchange of H⁺ for sodium (Na⁺) and potassium (K⁺), which again help mitigate the severity of the acidosis. The loss of bone mineral leads to bone fragility, which will increase the rate of bone fracture.

The level of circulating PTH generally increases as kidney function declines, termed secondary hyperparathyroidism.¹⁴⁰ Elevated PTH causes net calcium and phosphate release from bone via osteoblast-mediated activation of osteoclasts.¹⁴¹ Metabolic acidosis suppresses osteoblast-induced collagen synthesis, increases PTH secretion, and increases osteoclast activity, thus shifting the balance of bone formation and bone resorption toward increased bone turnover and net bone loss.^{36,99,136,142,143} Metabolic acidosis enhances the calcemic response of bone to PTH, possibly mediated by enhanced uptake of PTH by bone cells and by enhanced PTH-mediated cyclic adenosine monophosphate production.^142,144 The additive effects of acidosis and PTH were demonstrated in vitro with the mouse calvariae model; acidosis and PTH independently stimulated calcium efflux from bone, inhibited osteoblastic collagen synthesis, and stimulated osteoclastic β-glucuronidase secretion.¹³⁰ When calvariae were cultured in acidic medium with PTH, there was greater calcium efflux, less osteoblastic collagen synthesis, and more osteoclastic β-glucuronidase secretion than with either acid or PTH alone.¹³⁰

The level of the phosphaturic hormone, FGF23, increases significantly with decreasing kidney function, leading to decreased renal tubule reabsorption of inorganic phosphate and decreased 1,25 (OH)₂D₃ production.¹⁴⁵ The elevation of FGF23 is associated with an increased mortality.^146,147 FGF23 is synthesized in osteoblasts and osteocytes¹⁴⁸; however, the primary regulator of its synthesis is not clear.¹⁴⁹ Metabolic acidosis also develops and increases in severity as CKD progresses, suggesting the hypothesis that acidosis could stimulate FGF23 production. Indeed, in neonatal mouse calvarial organ cultures, physiological metabolic acidosis increased FGF23 gene expression and secretion of FGF23.¹⁵⁰ In primary calvarial osteoblasts, FGF23 gene expression was stimulated by acidosis within 6 hours, even in the absence of any additional phosphate that would have been released during acid-induced bone resorption.¹⁵⁰ Pharmacologic blockade of intracellular Ca signaling or cyclooxygenase 2 synthesis blocked the acidosis-induced FGF23 gene expression in primary osteoblasts and blocked acidosis stimulation of bone resorption,¹⁵¹ suggesting that the initial signaling responses to metabolic acidosis that stimulate FGF23 are the same that lead to increased bone resorption. Thus, metabolic acidosis could contribute to the increased FGF23 observed in patients with CKD and metabolic acidosis.

The 1,25 dihydroxyvitamin D is necessary for normal bone formation and resorption.¹⁵² Metabolic acidosis reduces vitamin D production in animals,^{55,57,83,153,154} and alkali correction of metabolic acidosis in patients with CKD increases serum 1,25 dihydroxyvitamin D levels, despite experimentally maintained serum levels of ionized calcium concentration and no changes in serum levels of magnesium, phosphate, albumin, or 25-hydroxyvitamin D.^155,156

It is difficult to isolate the independent effects of metabolic acidosis on the bone disease observed in CKD.^39,157,158 However, a retrospective study using bone biopsies to correlate bone health with acid-base status in patients with CKD revealed that subjects with a normal bone biopsy had a serum bicarbonate level within the normal range, whereas those with overt bone disease had mild-to-moderate acidosis.¹⁵⁹ The complex connection between bone health and acidosis is apparent in studies of chronic hemodialysis patients. In those patients with predialysis serum bicarbonate <20 mEq/L or >22 mEq/L, a higher incidence of bone fractures was observed.¹⁶⁰ In a randomized bone biopsy study of 21 hemodialysis patients, correction of metabolic acidosis prevented progression of hyperparathyroidism in those with high turnover and stimulated turnover in those with low bone formation.⁶¹ In a prospective 3-month study with oral bicarbonate or placebo given to 40 patients with mild-to-moderate CKD, correction of metabolic acidosis attenuated the increase in blood urea nitrogen and in PTH (nonsignificantly) but led to increased edema in 2 patients and the use of more antihypertensive agents in 3 patients.¹⁶¹

The clinical studies in which patients with CKD and metabolic acidosis were treated with oral alkali to neutralize systemic acidosis, dietary measures to reduce acid intake, or an experimental intestinal acid binder, which results in an increase in serum bicarbonate, primarily addressed treatment-dependent changes in glomerular filtration.^162–164 There are few studies in which the effects of treating metabolic acidosis examined changes in bone. Treatment of acidosis (n = 19) compared with control (n = 11) improved bone histology but not bone density in transplant patients.¹⁶⁵ A recent 24-month, multicenter, randomized, placebo-controlled clinical trial of 149 patients given either 0.4 mEq/kg of ideal body weight of sodium bicarbonate per day or a placebo did not find a significant difference in bone mineral density.¹⁶⁶ A pragmatic multicenter, parallel-group, double-blind, placebo-controlled randomized trial of 300 elderly acidic CKD patients did not find a difference in markers of bone mineral metabolism in the 187 patients with primary outcome data available after 1 year.¹⁶⁷ Their planned substudy of bone mineral density was discontinued because of low recruitment rates. The patients treated with bicarbonate experienced more clinical adverse effects. Correction of metabolic acidosis in patients with CKD and vascular calcification increases the possibility that the vascular calcification will worsen as the systemic pH increases, because calcium phosphate complexes that form the calcification are less soluble in an alkaline milieu.^168,169 Although this topic has not been extensively studied to date, there is no evidence in humans that treatment of metabolic acidosis worsens calcification,¹⁷⁰ and one prospective, randomized, crossover pilot study in 20 acidemic patients with CKD found that treatment of metabolic acidosis with sodium bicarbonate significantly improved vascular endothelial function.¹⁷¹ Future appropriately powered, long-term studies are needed to address the potential effect of treatment of metabolic acidosis in patients with CKD on bone mass and quality and vascular calcification.

Metabolic acidosis leads to an increase in bone resorption and a decrease in bone formation. One can hypothesize that the resorptive state of the bone mineral will directly affect the magnitude of the bone buffering.¹³⁹ Patients with adynamic bone disease, the most common form of renal osteodystrophy, have a marked decrease in the activity of both osteoblasts and osteoclasts.¹⁷² In patients with this disorder, although metabolic acidosis would be expected to stimulate osteoclastic resorption, it would not be expected to further decrease the already significantly suppressed osteoblastic bone formation, resulting in decreased proton buffering by bone. This would be expected to worsen the acidosis and serve to increase the adverse effects of acidosis on the kidney, muscle, and bone.¹ Bisphosphonate treatment in these individuals would further decrease proton buffering by bone.¹³⁹ Conversely, in high turnover renal osteodystrophy, there is an increase in the activity of both the osteoclasts and the osteoblasts. In this condition, there might be enhanced proton buffering during metabolic acidosis, mitigating the adverse effects of acidosis. Clinical studies of the magnitude of proton buffering by bone in patients with low and high turnover renal osteodystrophy are needed to explore this important topic.

Comparison of respiratory acidosis to metabolic acidosis

Bone responds far more robustly to an isohydric reduction in pH induced by a decrease in bicarbonate (metabolic acidosis) compared with an increase in the partial pressure of carbon dioxide (respiratory acidosis).^{19–23,93,101,102,111,173–176} In neonatal mouse calvariae, the type of acidosis is critical to determining the magnitude of both the bone calcium release and hydrogen ion buffering. During acute incubations (<3 hours), a model of metabolic acidosis induces far more calcium release than an isohydric model of respiratory acidosis.²¹ When osteoblasts are cultured to form bone nodules, metabolic, but not, respiratory acidosis inhibits bone formation.¹¹¹ During metabolic acidosis, there is hydrogen ion influx onto the bone surface, which does not occur during respiratory acidosis.²¹

As opposed to metabolic acidosis, cell-mediated net calcium efflux is not observed during chronic respiratory acidosis.¹⁰¹ Respiratory acidosis does not alter osteoblastic collagen synthesis, alkaline phosphatase activity, or osteoclastic β-glucuronidase release, or appreciably alter the surface ion concentration of bone.^93,106 Respiratory acidosis does not increase PGE₂ levels in neonatal mouse calvariae, which is critical for induction of osteoclastic bone resorption by receptor activator of nuclear factor-κB/RANKL.¹⁷⁷ The fact that metabolic, but not respiratory, acidosis adversely affects bone suggests that a low bicarbonate in addition to a low pH is critical for overall bone dissolution and resorption.⁸⁹

Conclusion

Metabolic acidosis induces changes in the bone mineral that are consistent with the role of bone mineral as a proton buffer. The decrease in mineral sodium, potassium, carbonate, and phosphate will each lead to protons being buffered and to an increase in systemic pH back toward the physiological normal. These changes in mineral composition come about first through physicochemical mineral dissolution and later through alterations in bone cell function. Through the proton receptor OGR1, acidosis suppresses osteoblastic collagen synthesis and stimulates prostaglandin production, which, acting in a paracrine manner, increases osteoblastic RANKL synthesis. RANKL then stimulates osteoclastic activity and recruitment of new osteoclasts to further promote bone resorption and buffering of the proton load. There is, in addition, a direct effect of acidosis on osteoclasts to induce bone resorption. This protective function of bone to maintain systemic pH, which has a clear survival advantage for mammals, comes at the expense of its mineral stores.