Glutamine energy substrate anaplerosis increases bone density in the Pahenu2 classical PKU mouse in the absence of phenylalanine restriction

Abstract Osteopenia is an under‐investigated clinical presentation of phenylalanine hydroxylase (PAH)‐deficient phenylketonuria (PKU). While osteopenia is not fully penetrant in human PKU, the Pahenu2 mouse is universally osteopenic and ideal to study the phenotype. We determined Pahenu2 mesenchymal stem cells (MSCs) are developmentally impaired in the osteoblast lineage. Moreover, we determined energy dysregulation and oxidative stress contribute to the osteoblast developmental deficit. The MSC preferred substrate glutamine (Gln) was applied to enhance energy homeostasis. In vitro Pahenu2 MSCs, in the context of 1200 μM Phe, respond to Gln with increased in situ alkaline phosphatase activity indicating augmented osteoblast differentiation. Oximetry applied to Pahenu2 MSCs in osteoblast differentiation show Gln energy substrate increases oxygen consumption, specifically maximum respiration and respiratory reserve. For 60 days post‐weaning, Pahenu2 animals received either no intervention (standard lab chow), amino acid defined chow maintaining plasma Phe at ~200 μM, or standard lab chow where ad libitum water was a 2% Gln solution. Bone density was assessed by microcomputed tomography and bone growth assessed by dye labeling. Bone density and dye labeling in Phe‐restricted Pahenu2 was indistinguishable from untreated Pahenu2. Gln energy substrate provided to Pahenu2, in the context of uncontrolled hyperphenylalaninemia, present increased bone density and dye labeling. These data provide further evidence that Pahenu2 MSCs experience a secondary energy deficit that is responsive both in vitro and in vivo to Gln energy substrate and independent of hyperphenylalaninemia. Energy support may have effect to treat human PKU osteopenia and elements of PKU neurologic disease resistant to standard of care systemic Phe reduction. Glutamine energy substrate anaplerosis increased Pahenu2 bone density and improved in vitro MSC function in the context of hyperphenylalaninemia in the classical PKU range.

Pah enu2 MSCs, in the context of 1200 μM Phe, respond to Gln with increased in situ alkaline phosphatase activity indicating augmented osteoblast differentiation. Oximetry applied to Pah enu2 MSCs in osteoblast differentiation show Gln energy substrate increases oxygen consumption, specifically maximum respiration and respiratory reserve. For 60 days postweaning, Pah enu2 animals received either no intervention (standard lab chow), amino acid defined chow maintaining plasma Phe at $200 μM, or standard lab chow where ad libitum water was a 2% Gln solution. Bone density was assessed by microcomputed tomography and bone growth assessed by dye labeling. Bone density and dye labeling in Phe-restricted Pah enu2 was indistinguishable from untreated Pah enu2 . Gln energy substrate provided to Pah enu2 , in the context of uncontrolled hyperphenylalaninemia, present increased bone density and dye labeling. These data provide further evidence that Pah enu2 MSCs experience a secondary energy deficit that is responsive both in vitro and in vivo to Gln energy substrate and independent of hyperphenylalaninemia. Energy support may have effect to treat human PKU osteopenia and elements of PKU neurologic disease resistant to standard of care systemic Phe reduction. Glutamine energy substrate anaplerosis increased Pah enu2 bone density and improved in vitro MSC function in the context of hyperphenylalaninemia in the classical PKU range.

K E Y W O R D S
glutamine, osteopenia, oxidative phosphorylation, Pah enu2 , phenylketonuria
Phenylketonuria (PKU) osteopenia was identified in the 1960s and clinical description of osteopenic PKU patients is extensive. [8][9][10][11] Lumbar spine bone mineral density Z scores of À2.0 are observed among early-identified, continuously treated patients. 12 Similar reduction in total body bone mineral density is observed. 13 Equivalently low bone mineral density occurs in therapy noncompliant patients. 12 Pathophysiological mechanisms of PKU bone disease remain ambiguous. The bone phenotype was originally attributed to diet therapy where by an undefined mechanism, bioavailability of calcium, phosphorous, and other bone-forming material were reduced; however, this is poorly supported as osteopenia is recognized in patients that never received diet therapy and young patients following short-term therapy. 12 Several studies find no correlation 8,11,[14][15][16][17] between hyperphenylalaninemia and bone disease; others show a negative correlation between hyperphenylalaninemia and bone disease. 28,29 Biochemical ambiguity does not end with Phe homeostasis as representation of bone formation markers, 18,19 bone resorption markers, 20,21 and other metrics related to bone provide no means to inform osteopenia risk. 18,20,22 Phenylketonuria osteopenia has knowledge gaps as investigation is descriptive and pathophysiology is largely un-investigated. While the bone phenotype is not fully penetrant in humans, osteopenia is universal in the Pah enu2 mouse model of classical PKU. Our previous investigations of Pah enu2 osteopenia identified a mesenchymal stem cell (MSC) developmental defect involving energy deficit and oxidative stress. [23][24][25] These investigations were the first to assess bone development as a participatory element in the pathology of the PKU osteopenia. Here, we apply the MSC preferred energy substrate glutamine to in vitro MSC differentiation and mitochondrial oxygen consumption. [26][27][28][29] Gln enhances MSC osteoblast development and increases mitochondrial oxygen consumption. In vivo a post-weaning Gln regimen increased Pah enu2 bone density. Moreover, augmentation of in vitro MSC metrics and in vivo bone density was achieved within the context of hyperphenylalaninemia in the classical range. Energy repletion provides an alternative intervention to treat Pah enu2 osteopenia acting independent of systemic Phe homeostasis.

| Pah enu2 and control animals
Pah enu2 and C57bl/6 were propagated at the Rangos Research Center animal facility at Children's Hospital of Pittsburgh with an approved protocol. Pah enu2 and control animals were generated matings heterozygous females and heterozygous males. Offspring were genotyped as described. 30 Alternative homozygous genotypes (experimental enu2/enu2, control wt/wt) were used in experimental cohorts. After weaning (day of life 21), Pah enu2 animals were provided one of the following diets 1. Standard mouse chow; 2. standard mouse chow where the ad libitum water supply was a 2% Gln solution; 3. Phe-free amino acid defined chow with Phe supplemented in drinking water (0.35 g/L). Standard chow and standard chow plus 2% Gln maintain Pah enu2 plasma of $2000 μM in males and 2200 μM in females. Phe-free amino acid defined chow produces Pah enu2 plasma Phe of $200 μM. 30 Animals provided dietary Phe restriction were utilized in in vivo histomorphometry studies. Owing to Gln aqueous instability, over the 60-day post-weaning regimen, a freshly dissolved 2% Gln solution was provided each Monday, Wednesday, and Friday. Control (wt/wt)

Highlights
• Pahenu2 osteopenia involves energy deficit. • Glutamine energy substrate anaplerosis increases Pahenu2 mesenchymal stem cell functionality. • An in vivo glutamine regimen increased Pahenu2 bone density independent of hyperphenylalaninemia.
littermates were provided standard mouse chow. Experiments used animals in the fed state. Animals (control, Pah enu2 ) were sacrificed by CO 2 asphyxiation, at 2-3 months of age.

| Mesenchymal stem cell culture and osteoblast differentiation
Experimental and control cohorts contained at least four animals (equal male/female representation). MSCs were prepared from 2-to 3-month animals (control, Pah enu2 ) as described. 23 Over the course of differentiation, media were replaced every Monday, Wednesday, and Friday. MSC osteoblast differentiation was assessed by in situ alkaline phosphatase activity and mineralization. In situ alkaline phosphatase activity used 0.01% napthol AS-MX substrate and fast blue product visualization as described. [23][24][25] Visualizing mineralization used von Kossa silver staining as described. [23][24][25] Densitometry analysis was performed after converting the image to grayscale, inverted, and the mean white area was measured with ImageJ software.

| Oximetry of MSC cultures in osteoblast differentiation
Proliferating Pah enu2 and control MSCs (40 000 cells) were plated in 96-well Seahorse oximetry plates. At confluence, osteoblast differentiation was induced as above. Experimental Pah enu2 cultures were provided media including 1% Gln. On day 14 post-induction, 20 h prior to assessment, differentiating MSCs were washed and unbuffered media provided. Experimental conditions (supplemental Phe, Gln) were maintained in unbuffered media. Assessment utilized the Seahorse Mito Stress Test that applies oligomycin, carbonyl cyanide-4-phenylhydrazone (FCCP), 2-deoxy-glucose, and rotenone/antimycin to determine basal respiration, maximal respiration, ATP production, and spare capacity. 23,32,33 During assessment, each experimental and control condition utilized eight replicate wells. Data analysis is as described using Graph pad software and unpaired T-test. 23,32,33 F I G U R E 2 Gln substrate rescues mitochondria oxygen consumption in differentiating Pah enu2 MSCs in hyperphenylalaninemia FIGURE 2. Assessment of differentiating MSC from six Pah enu2 preparations (3 male and 3 female) and Six C57bl/6 (3 male and 3 female). Gln support of WT MSCs did not change respiration (data not shown); OCR, oxygen consumption rate. **** p ≤ 0.0001; *** p ≤ 0.001 F I G U R E 1 Gln increased in situ alkaline phosphatase activity in hyperphenylalaninemia FIGURE 1. MSCs from Pah enu2 (2 male and 2 female) and C57bl/6 (2 male and 2 female) were differentiated in standard media or media supplemented with 1% Gln. *p ≤ 0.05; ****p ≤ 0.0001

| Microcomputed tomography and dye labeling
Experimental (Pah enu2 Phe restricted diet, Pah enu2 Gln energy substrate diet) and control (C57bl/6, unmanaged Pah enu2 ) animal cohorts (minimum 6 animals, equal male vs. female representation) were sacrificed 2 months post-weaning. Trabecular bone analysis applied microcomputed tomography as described in a blinded manner. 24,25 Briefly, fixed lumbar vertebrae, were scanned in 70% ethanol at 6-μm resolution with a 0.25-mm aluminum filter; voltage and current were set at 69 kV and 100 μA. Image reconstruction utilized Nrecon and InstaRecon software. Cross sectional images used Dataviewer software. Quantitative analysis assessed a 1.2-mm region at the midpoint of the fourth lumbar vertebrae. Determined were bone volume/total volume, bone surface density, trabecular number, and total porocity. Data represent that of the entire cohort (experimental, control) not parsed by sex. Dynamic histomorphometry applied 7 days and 2 days prior to sacrifice intra-peritoneal injection of calcein (20 μg/g mouse weight) and xylenol orange (80 μg/g mouse weight), respectively. Cutting undecalcified 7 μm frozen sections applied tape means. Measurement of dye-labeled surface was performed as described. 24 3 | RESULTS 3.1 | In the context of hyperphenylalaninemia, Gln energy substrate increased Pah enu2 in situ alkaline phosphatase activity Figure 1 provides staining of in situ alkaline phosphatase activity. As previously reported, Pah enu2 MSCs in osteoblast differentiation have decreased alkaline phosphatase activity compared to C57bl/6. Pah enu2 MSCs provided Gln energy support demonstrate a statistically significant increase in in situ alkaline phosphatase activity. Mineralization did not increase with Gln energy substrate (data not shown).

|
In the context of hyperphenylalaninemia, Gln energy substrate increased Pah enu2 MSC oxygen consumption in osteoblast differentiation Figure 2 oximetry applied the Seahorse Mito Stress Test to MSCs following 14 days of osteoblast differentiation. Basal oxygen consumption and ATP production are equivalent between C57bl/6, Pah enu2 , and Pah enu2 provided Gln energy substrate. Maximum respiration and spare capacity of Pah enu2 and Pah enu2 with 1% Gln are both less than that of C57bl/6 controls. However, Gln energy substrate provided Pah enu2 MSCs statistically significant increases to both maximum respiration and spare capacity.

| In vivo Gln energy substrate increases Pah enu2 bone density in uncontrolled hyperphenylalaninemia
In vivo assessment provided Pah enu2 Gln energy substrate from weaning (day of life 21) continuously for the following 60 days ad libitum as a 2% Gln solution in the water supply. These animals otherwise received normal chow causing unregulated Phe homeostasis of $2000 in males and 2200 μM in females. For comparison, a separate Pah enu2 cohort was provided a Phe restricted diet, maintaining blood Phe of $2200 μM. Microcomputed tomography, assessing the fourth and fifth lumbar vertebrae, determined bone density of untreated Pah enu2 and Phe-restricted Pah enu2 is identical. All metrics (bone volume/total volume, bone surface density, trabecular number, and total porocity) are indistinguishable and significantly less than unaffected control littermates. Pah enu2 provided Gln energy substrate anaplerosis, in the context of unrestricted Phe homeostasis, show improved static histomorphometry metrics and improved dye labeling. Bone density improved among Phe unrestricted Pah enu2 receiving Gln energy substrate anaplerosis. Similarly, calcein and xylenol orange dye labeling was identical between control Pah enu2 and Phe-restricted Pah enu2 . Gln substrate increased dye labeling to equivalence with unaffected littermates.

| DISCUSSION
Osteopenia is an under-investigated phenotype in PAHdeficient PKU and pathophysiology remains ill-defined. PKU clinical phenotypes occur in brain and bone. As neither tissue express the PAH gene nor hydroxylate Phe, clinical phenotypes arise from secondary effect of biochemical dysregulation. The universally osteopenic Pah enu2 mouse is an ideal model to study osteopenia as the phenotype is not fully penetrant in patients. We determine Pah enu2 MSCs are deficient in osteoblast differentiation where energy deficit and oxidative stress are contributing factors. [23][24][25] In the context of hyperphenylalaninemia, the MSC preferred energy substrate Gln increased in vitro alkaline phosphatase activity (a measure of osteoblast differentiation) and mitochondrial oxygen consumption (a measure of oxidative energy production). In vivo Gln energy substrate improved bone density concurrent to unregulated hyperphenylalaninemia. While resting and proliferating MSCs are glycolytic, those in the course of osteoblast differentiation require oxidative phosphorylation. Respirometry applied to mitochondria from Pah enu2 MSC during osteoblast differentiation, showed an attenuated complex 1 response to pyruvate substrate; however, complex 1 response to glutamate (Glu) substrate was similar to controls. 23 Respirometry in Pah enu2 brain tissue mitochondria showed similar attenuated response to pyruvate substrate. 34 Pyruvate enters mitochondria through voltage dependent anion channels. The Phe catabolite phenylpyruvate is structurally similar to pyruvate and a recognized inhibitor of pyruvate mitochondrial transport. [35][36][37][38] Pyruvate transport inhibition reduces acetyl-CoA that drives Kreb cycle processivity and creation of reducing equivalents for oxidative phosphorylation. On the plasma membrane, MSCs express SLC1A5 high-affinity Gln transporter. Cytosolic glutaminase converts Gln to Glu. Two related proteins, the mitochondrial glutamate carriers GC1 (SLC25A22) and GC2 (SLC25A18) transport Glu into the mitochondria, where it is converted to alpha-ketoglutarate distally repleting the Kreb cycle. We posit Gln energy substrate increases Pah enu2 bone density by alternative pathway Kreb cycle repletion circumventing phenylpyruvate inhibition of pyruvate transport. Gln up-regulates energy homeostasis enabling osteogenesis (Figures 1-3). Augmented MSC function and increased bone density within the context hyperphenylalaninemia is a consequence of alternative energy pathway utilization. Moreover, realizing increased bone density in uncontrolled hyperphenylalaninemia strongly argues against causation by asymmetric amino acid transport through the LAT1 (SLC7A5 gene product) transporter.
Increased bone density is the first physiologically quantifiable PKU intervention response occurring independent of systemic Phe management. As most adolescent/adult PKU patients are therapy noncompliant, Gln energy support may provide a means to treat osteopenic patients unwilling/unable to engage systemic Phe reduction. Further, even more importantly, these data suggest energy deficit may contribute to PKU neurologic phenotypes. Even that small minority of adult/adolescent PKU patients that remain compliant to Phe reduction therapy present cognitive decline, executive function deficit, and other late onset neurologic phenotypes. Should energy deficit contribute to neurologic disease, energy support, be it Gln or other alternative pathway substrates, may provide an under-appreciated intervention opportunity. 39 Literature Cited.
National PKU Alliance to SFD.

CONFLICT OF INTEREST
No authors have competing interests related to these studies.
DATA AVAILABILITY STATEMENT Data will be made available upon reasonable request.

ETHICS STATEMENT
No patients were involved in these studies.

INSTITUTIONAL COMMITTEE FOR CARE AND USE OF LABORATORY ANIMALS
Pah enu2 and C57bl/6 animals are managed under an approved protocol of the Children's Hospital of Pittsburgh IACUC.