Examining the blood amino acid status in pretherapeutic patients with hyperphenylalaninemia

Abstract Background Hyperphenylalaninemia is the most common genetic metabolic disease. Early treatment prevents brain injury effectively. The present study aimed to detect the exact amino acid status of patients with hyperphenylalaninemia before treatment. Methods Data of 116 newborn patients from our Newborn Screening Center and 161 older patients from our clinic before treatment were collected. The content of 17 amino acids in their blood was determined by tandem mass spectrometry and compared with normal controls. Relationship between phenylalanine and other amino acids in patients was analyzed using the smoothing curve fitting and threshold effect analysis. Results Most amino acids in the blood of patients were within the normal range; however, they were different significantly from those of the normal children. Newborn patients showed higher phenylalanine (346.30 vs 45.90 µmol/L), valine (121.50 vs 110.30 µmol/L), citrulline, ornithine and lower tyrosine (52.97 vs 66.12 µmol/L), threonine (68.68 vs 78.21 µmol/L), glutamine levels than observed in normal newborns. Older patients showed significantly higher phenylalanine (844.00 vs 51.82 µmol/L), valine (117.60 vs 110.90 µmol/L), histidine, serine and lower tyrosine (55.97 vs 67.31 µmol/L), threonine (35.94 vs 51.89 µmol/L), alanine, asparagine, glutamic acid, methionine, arginine, glycine, ornithine, glutamine content than found in matched normal children. Tyrosine, valine, ornithine, and threonine in newborn patients and tyrosine, glycine, glutamine, and threonine in older patients had a nonlinear correlation with phenylalanine levels with obvious threshold effect and clear inflection points. Conclusion Significant difference was observed in the amino acid status between pretherapeutic hyperphenylalaninemia patients and normal children. Some amino acids showed notable threshold effect with phenylalanine level in a nonlinear pattern.


| INTRODUC TI ON
Amino acid is the basic nutrient of the human body and the structural unit of proteins. Both proteinogenic amino acids and nonproteinogenic amino acids play an important role in biosynthesis, energy supply, and neurotransmission. For example, tyrosine (Tyr) and its precursor phenylalanine (Phe) are precursors of the catecholamine neurotransmitters dopamine, epinephrine, and norepinephrine and various trace amines.
Glycine (Gly) is a precursor of porphyrins such as heme. Glutamine (Gln) is the precursor of glutamic acid (Glu), and Glu is the main excitatory neurotransmitter in the human brain. 1 Arginine (Arg) is a precursor of nitric oxide. 2 Aspartic acid (Asp), serine (Ser), histidine (His), Gly, and Gln are precursors of nucleotides. 3 Tryptophan (Trp) is a precursor of the neurotransmitter serotonin. 4 Nonproteinogenic amino acids in humans, including gamma-amino-butyric acid (GABA), citrulline (Cit), and ornithine (Orn), are key metabolic intermediates, such as in the biosynthesis of the neurotransmitter and so on. For example, Orn and S-adenosylmethionine are precursors of polyamines. 5 In the clinical setting, multiple amino acids can be determined in dried blood spots by tandem mass spectrometry (MS/MS). Changes in amino acid profile may be found in patients with various inborn errors of metabolism. Hyperphenylalaninemia (HPA) is the first disorder to benefit from newborn screening (NBS). HPA is a metabolic condition characterized by high serum Phe levels, caused mostly by phenylalanine hydroxylase (PAH) deficiency and seldom by tetrahydrobiopterin (BH4) deficiency. 6 A recent recognition of the DNAJC12 deficiency, which is the co-chaperone of PAH, expanded the clinical and metabolic spectrum of HPA. 7 The term HPA should encompass all of the above conditions. PAH deficiency, also called phenylketonuria (PKU, OMIM no. 261600) and accounting for 90% of HPA cases, is the most frequent inherited amino acid metabolic disease with the incidence of approximately 1:11 000 in China, leading to the accumulation of Phe and hence brain injury. A Phe-restricted diet is the primary effective therapy to prevent neurological damage in patients with PKU. 8, 9 HPA is marked by a serum Phe level of >120 µmol/L and serum Phe/Tyr level of >2. 10 HPA was classified according to the following cut-off values of Phe concentrations at the time of testing: mild (non-PKU) HPA (blood Phe: 120-600 µmol/L), mild PKU (blood Phe: 600-1200 µmol/L), and classical PKU (blood Phe: >1200 µmol/L). 11,12 In the 1960s, PKU became the first disease identified through the universal NBS program. Today, NBS in Shanghai is performed by MS/MS, which detects multiple amino acids besides Phe.
Although strict dietary adherence prevents profound neurologic injury, subtle cognitive deficits remain persistent in some patients who were treated early and received continuous treatment. 8 Myelin abnormality, which is marked by impairment in the white matter integrity and deficiencies in the dopamine levels, may also involve abnormal amino acid metabolism. 13,14 However, this finding cannot fully explain the general disturbance of neuronal function and other underlying convergent mechanisms.
With regard to the amino acid metabolism in PKU, most previous studies focused on the "large neutral amino acid (LNAA) hypothesis." LNAAs include Phe, valine (Val), Gln, threonine (Thr), Tyr, His, methionine (Met), leucine (Leu), Trp, and isoleucine (Ile). They are transported to the brain by the same amino acid transporter in the blood-brain barrier. In PKU, elevated plasma Phe levels impair the transport of other LNAAs, resulting in cerebral depletion of LNAA and hence decreased biosynthesis of protein and impaired brain function. 8, 15 Increased cerebral Phe competes with Gly and Glu for their receptors, changing the levels of receptor expression and thus influencing the synaptic function. 16 A previous study reported that the glutamatergic synaptic transmission in the brain is impaired in patients with PKU. 17 PKU patients have reduced myelin formation and dopamine levels. 18 However, only a few studies evaluated the serum amino acid status of PKU patients before treatment. The present study aimed to determine whether the status of other amino acids, except Phe, in patients with HPA is disturbed prior to treatment.
To clarify the amino acid status of HPA patients before treatment, we reviewed the amino acid spectrum of 277 patients in the past 15 years. We compared the patients' amino acid spectrum with that of normal children and got some interesting results. Then, we analyzed the correlation between amino acid levels and Phe at different concentrations. As a result, significant results were also exposed.

| Study population
Since the normal reference ranges of amino acids differ with age, we showed reference range of amino acids level for newborns and older children separately. The study participants were divided into four groups. We reviewed the NBS program in our screening center from 2003 to 2018 and collected the data of 116 HPA newborns; their blood samples were usually collected on the 4th day after birth after Conclusion: Significant difference was observed in the amino acid status between pretherapeutic hyperphenylalaninemia patients and normal children. Some amino acids showed notable threshold effect with phenylalanine level in a nonlinear pattern.

K E Y W O R D S
amino acids, hyperphenylalaninemia, newborn screening, phenylalanine hydroxylase deficiency, phenylketonuria fully feeding for 3 days (group 1). A total of 150 healthy newborn controls were selected randomly from the NBS database in the same birth period with newborns in group 1 (group 2). Another 161 HPA patients, aged 1 month to 5 years old, who were suspected to have HPA by other hospital, transferred to our clinic at several months of age mostly, and finally confirmed with HPA in our center from 2003 to 2018 were selected and comprised group 3. All the patients seemed stature symmetrical with normal weight and height when they visited our clinic. All the registered body mass index (BMI) values were within normal range according the WHO standards. Those with other accompanying chronic diseases were excluded. A total of 200 blood spot samples from age-matched healthy children were used as control (group 4). Blood samples were collected while the patient was in a fasting state, and the samples detected within 48 hours after blood collection. The patients' screening results were analyzed and compared with their respective controls. Furthermore, the correlation between the differential amino acid levels and Phe concentrations was also analyzed in different age groups.

| Analytical methods
The amino acids in dried blood pots on filter paper were determined by MS/MS (Applied Biosystems, API 2000). All the chemicals used in MS/MS (methanol, and acetonitrile) were liquid chromatographic/ mass purity grade and were purchased from Fisher Scientific. The preparation and detection methods were described in previous studies. 19,20 In brief, dried blood spots were extracted using methanol, which contained amino acids as internal standards. After derivatization with n-butyl alcohol hydrochloric acid, the samples were tested using MS/MS. A total of 17 amino acids were determined, including Phe, Tyr, alanine (Ala), Asp, Glu, Met, Val, Arg, Gly, Gln, His, Ser, Thr, Leu, Trp, Cit, and Orn.

| Statistical analysis
The results were expressed as arithmetic mean ± standard error of the mean (mean ± SEM). T-test was used to study the differences in the concentrations of amino acids. Statistical analyses were performed using the SPSS version 22 statistical software (SPSS Inc). The correlation between the differential amino acid levels and Phe concentrations was analyzed using the EmpowerStats software. Using smoothing curve fitting, we determined whether a specific amino acid and Phe have linear or nonlinear correlation. We used Pearson's regression to determine the linear correlation between variables and threshold effect analysis for curve correlation. A P value of <.05 was considered significant.

| Ethics statement
The study protocol was approved by the Ethics Committee of Xin Hua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (XHEC-D-2014-141). Parents were properly informed about the study and signed a written consent.

| Baseline characteristics of patients and controls
The characteristics and Phe concentrations of the four groups are summarized in Table 1   They were assigned as the control of group 3, with normal Phe, and were age-and sex-matched with group 3.

| Pretherapeutic amino acid status in HPA patients
The blood amino acid concentrations in the four groups are summarized in  Figure 2).

| Correlation between Phe and the different amino acid levels in patients
The correlation between Phe and these differential amino acid lev-     F I G U R E 2 Differences in amino acid levels between older HPA patients and controls. Data are presented as mean ± SEM. **P < .01; *P < .05  To summarize the general impression of the patients in the two groups, higher Phe was associated with lower levels of Tyr, Thr, Orn, Gly, and Gln in a nonlinear pattern. Moreover, different threshold Phe concentrations existed in their respective smooth curves.

| D ISCUSS I ON
The present study focused on the initial amino acid status of pretherapeutic HPA children, early to several days after birth. Although all the other amino acids determined, except for Phe, were within the normal ranges, many amino acids were significantly different from those of the normal group. In addition to higher Phe and lower Tyr in HPA patients, Using the threshold effect analysis, we obtained a general impression that higher Phe concentration was associated with lower Tyr, Thr, Orn, Gly, and Gln levels in a nonlinear pattern. Most of the involved amino acids were LNAA. This finding indicated that Phe produces a pathological effect through LNAA even before treatment.
In a recent study by Suat Ekin on PKU patients, with a mean age of 6.92 ± 0.61 years (n = 30), a tendency of low levels of serum Ala, Gly, and Orn was exposed in patients, although no statistical difference was observed between the PAH deficiency group and healthy group. 21 In our study, the older HPA children showed significantly lower levels of Tyr, Ala, Gly, Orn, Thr, Asp, Glu, Met, Arg, and Gln than in the corresponding control group, all with statistical difference. In this point, the results of the two studies were mutually supportive. A statistical difference was observed in our study, while the previous study showed otherwise; we attributed this finding to the sample size used in this study (n = 277), which was larger than that of the previous study (n = 30).
Tyr is synthesized from Phe for normal person. However, the synthesis of Tyr from Phe is impaired in PKU, resulting in lower Although within normal range, Gln in the blood of both group 1 and group 3 was significantly lower than its respective control groups. Gln is the precursor of Glu, which is the main excitatory neurotransmitter in the human brain. 1 Just like Thr, Gln is also one of the LNAAs competing with Phe in the process of transporting to the brain; we speculated that the decrease of Gln in blood may lead to its weaker competition with Phe to get into the brain, which can lead to poor glutamate synthesis in the brain. Glu levels in group 1 were within normal range and were similar to that of group 2. However, in older groups, Glu level in group 3 was obviously lower than that in group 4, although the value was still within normal range. The relative lack of Glu might further weaken their competitiveness as neurotransmitters. This finding is similar to that of a previous study and showed that glutamatergic synaptic transmission is impaired in the brain of PKU patients. 1,17 The relative decrease of Gln and Glu in the blood exposed in our study might be one of the causes of this impairment or may aggravate it.
Arg is a semi-essential factor because it may not be synthesized sufficiently in growing children. 22 In our study, Arg level in group 3 was lower than that in group 4, while Arg level in group 1 was similar to that in group 2. These results indicated that more and more Thr levels in both group 1 and group 3 were significantly lower than the levels of respective control groups. Thr can convert to pyruvate and produce Gly and acetyl-CoA, which are the most

F I G U R E 4
The smooth curve showing the nonlinear correlation between different amino acid levels in group 3. The four smooth curves represent the correlation between Phe and Tyr, Thr, Gln, and Gly, respectively. Red points: smooth curve of amino acid; blue points: 95% confidence interval important intermediate molecules in energy metabolism. Thr is an essential amino acid. Decrease in Thr levels might lead to poor protein synthesis and energy metabolism disorder. Since Thr is one of the LNAAs competing with Phe in the process of transporting to the brain, we speculated that a decrease in serum Thr levels may lead to its weaker competition with Phe, and thus only lesser amounts of Thr will be transported into the brain, leading to poor protein synthesis in the brain. However, the mechanism behind the reduction in serum Thr levels of HPA patients remains unclear. In our study, group 3 had lower Gly levels (186.30 ± 3.87 µmol/L) than group 4 (247.00 ± 5.17 µmol/L), with the changes consistent with Thr. This finding is possibly due to the fact that Thr can be converted to Gly.
Along with the increase in Phe, both Thr and Gly showed the same trend; that is, they initially decrease and then eventually increase ( Figure 3). The consistency of trends observed for Thr and Gly fully confirms this point.
Metabolic syndrome including obesity and insulin resistance is increasing in PKU due to the following factors: diet and the disease itself. Kanufre demonstrated that PKU patients with excess weight showed higher blood concentrations of basal insulin. 23 Val is associated with insulin resistance: higher levels of Val are observed in the blood of patients with insulin resistance. Kara R. Vogel found that the serum Val levels of PKU mice were elevated. 24 Similarly, in our study, the levels of Val in group 1 and group 3 were significantly higher than those of their respective control groups. This phenomenon in our study supported the insulin elevation in PKU patients reported in the previous study. We speculated that the elevation of Val might have a correlation with the insulin levels in HPA patients, which requires further study to confirm this result.
Ala is closely related to glucose metabolism. In mammals, Ala plays a key role in glucose-Ala cycle, which is important for gluconeogenesis and energy supply. Besides, pyruvate kinase is the key enzyme in the process of glycolysis. The previous study has indicated that Phe inhibits pyruvate kinase and decreases glycolysis and energy production. Ala, a known competitor of Phe on the enzyme activity, prevents the inhibitory effect of Phe on the enzyme activity and prevents the reduction of glycolysis and energy production caused by Phe. 25,26 In our study, group 3 had significantly lower Ala levels than controls. We speculated that the decreased level of Ala might lead to a weaker protective function of glycolysis and energy production against Phe for HPA patients. Further study is needed to confirm this finding.
The current study did not include the comprehensive mechanism, but only described the amino acid spectrum in HPA patients before treatment. Overall, several variations were observed in the amino acid spectrum of pretherapeutic HPA children. Disturbance of amino acids in older patients was more serious than that in newborn patients. Compared with their controls, the relative lack of Thr, Ala, Asp, Glu, Met, Arg, Gly, Orn, and Gln in the blood of patients suggested that the nutritional problems might have occurred before the diagnosis and treatment. These findings might provide new insights into the metabolism of brain injury in HPA.
In conclusion, this study suggests that although the levels of amino acids (except Phe) in pretherapeutic HPA patients remained normal, significant differences existed already in many amino acids, compared with their respective controls. HPA patients before treatment had different amino acid spectrum even at neonatal period.
The pretherapeutic amino acid spectrum of patients varied obviously with age. In HPA neonatal patients, the levels of some amino acids were higher, while those of other amino acids were lower than the amino acid levels in normal newborns. As the patients' age increased, levels of more amino acids decreased, including Tyr, Thr, Ala, Asp, Glu, Met, Arg, Gly, Orn, and Gln. Some amino acids, especially several LNAAs, showed notable threshold effect with Phe level in a nonlinear pattern. The disturbed pretherapeutic amino acid status might partly be contributed to the occurrence of brain injury in PKU patients. Our findings indicated that more attention should be paid to the nutritional status of HPA patients as early as possible.