Fifteen-year follow-up of Italian families affected by arginine glycine amidinotransferase deficiency

Arginine:glycine amidinotransferase (AGAT, OMIM 602360) deficiency (AGAT-d) is a very rare inborn error of creatine (Cr) synthesis described in 2000 in an Italian family and successively confirmed by molecular and enzymatic analysis [1, 2]. This disorder is caused by a deficiency of the first enzyme involved in Cr synthesis (Fig. 1), resulting in a commonly recognized biochemical pattern represented by low Guanidinoacetic acid (GAA) concentrations in plasma and urine and low/undetectable brain Cr detectable by magnetic resonance spectroscopy (MRS). Main clinical symptoms of this disorder are intellectual disability, severe language impairment and behavioural problems. Similar to guanidinoacetate-methyltransferase deficiency (GAMT-d), another Cr synthesis defect, it is susceptible of supplementation therapy with high doses of Cr monohydrate per os [3‐6].

Since the first reports on two Italian families [1, 2, 6] 15 years ago, only 16 patients from eight families of different ethnic origins have been diagnosed [7].

There is a substantial lack of information on the outcome of these patients, due to recent detection of disease, limited number of years of substitutive treatment and paucity of identified patients. So far, a longitudinal 9-year study on a AGAT patient [8] and a recent review with some data on all AGAT-d patients identified worldwide [7] have been reported. In particular, this review confirms that the main clinical symptoms of the disease are present in all patients and reveals that myopathic symptoms are also present in about half of them, particularly in patients with a late diagnosis (from 6 to 23 years). Seizures and behavioural problems have also been reported in a few cases [7]. Chronic Cr supplementation of varying dosages, ranging from 100 to 800 mg/Kg/day, restores brain Cr to 60–95% of normal in all patients monitored by MRS, improving cognitive and motor symptoms [7], with the exception of one patient in which it increased only to 18% of normal even after 7 years of supplementation [7, 9]. Since blood and cerebral Cr concentrations during treatment have not been reported in all patients, it is not possible to establish a correlation, if any, between brain and blood levels. This information would be interesting in order to tailor treatments for each patient, thus avoiding high dosages and potential metabolic dysregulation. Therefore in this paper, we report the clinical data observed in our AGAT-d patients over a long period.

In fact, Cr treatment in these patients was pioneering and required constant adjustments according to clinical and biochemical evidence. During treatment, we attempted to answer the following questions which arose over time relative to patient outcome: a) which is the best dosage regimen? b) how long are brain Cr recovery times? c) how is neuropsychological impairment modified over time? d) is long term treatment safe for patients?

We updated previous data collected over time from the first Italian patients [1, 2, 6, 10‐12] and we reviewed their long term outcomes in relation to clinical, neuropsychological, biochemical and toxicological findings after 10/15 years of Cr supplementation per os.

It has been well recognized that Cr supplementation is able to restore brain pool in patients with AGAT-d but unfortunately little information is available in terms of dosage, clinical progression and disease outcome. This study aimed at providing an update of results achieved in 4 AGAT-d patients after 15 (P1-P3) and 10 years (P4) of Cr supplementation, respectively.

Our experience tells us that the first critical step in diagnosis of AGAT-d consists of a proper quantitative measure of Cr and GAA in body fluids. As reported in the first paper by our group [1], misleading results could arise from urine samples, where levels of two key metabolites were near the lower end of normal value range, probably reflecting excretion of Cr assumed in food. Improvements in diagnostic techniques have eliminated this problem, and successive measurements in the same patients demonstrated that Cr and GAA levels in AGAT-d were below control range both in blood and urine indicating that they are, together with cerebral Cr levels, the best markers for AGAT-d.

The decision to set the initial dose at 400 mg/Kg/day in five administrations was based on an adopted schedule in a patient with GAMT deficiency, the first described inborn error of Cr metabolism [19], where 50% replenishment of brain Cr pool after 12 weeks of supplementation was reported. Similar brain Cr restoration was observed in our patients after 9 months of therapy, which resulted in 80% of normal [1] in gray matter and cerebellum, with concomitant improvements in cognitive development and other clinical symptoms.

Similar ameliorations have been observed in all the other patients [8, 9, 20]; for one of them the dosage was increased up to 800 mg/Kg/day and then reduced to 500 mg/kg/day without different replenishment of brain Cr, while progress in cognitive development was noticed after start of treatment [8].

Even though all patients showed clinical improvements under Cr supplementation; in patients where treatment started at a later age, an amelioration in muscle strength/endurance [9, 20] and in adaptive functioning [9] was limited, while cognitive and speech remained unchanged. Taken together, these two findings support the hypothesis that myopathy may be a clinical sign of disease only at an older age [7, 20] and that irreversible brain damage occurs when treatment is delayed. Our data seem to confirm this hypothesis because all patients diagnosed at a young age did not show muscle deficiency and cognitive recovery was lower for patients older at diagnosis. In fact, TIQ of P1 did not change despite brain Cr recovery. Conversely, early treatment prevents adverse developmental outcomes, as demonstrated in P4 [10, 12] and in the other young patient who started treatment at 16.5 months [8]; both of them, currently 11 and 12 years of life respectively, have achieved normal development with excellent educational milestones. According to previous assessments [12], WISC-IV score in the asymptomatic patient was within normal range, indicating that his intellectual development proceeded typically overtime with a slight residual speech impairment, also present to a greater degree in his sisters.

The placental Cr delivery to the foetus during pregnancy is important for an adequate growth and a development in normal conditions [21, 22], and crucial in Cr disorders, as demonstrated by the normal clinical presentation at birth of P4. Maternal supplementation during pregnancy could have provided higher levels of Cr in the affected newborn, but she refused any prior treatment as well as any prenatal diagnosis. In order to correct Cr depletion, since the infant was breastfed, we initially tried to supplement the mother’s diet with Cr. Although this approach resulted in an increase of Cr concentration in maternal milk, Cr increase in the infant’s blood, urine and brain was unremarkable [10].

Considering the clinical results in our 4 patients and taking into account the limitations of this study, some recommendations based on the evidence might be useful in the management of AGAT-d patients. Daily dose was reduced initially to 200 mg/kg/day in P1 and P2 after 6 years, and in P3 after 5 years, respectively, with good results [11]. Subsequently, based on MRS data obtained in the youngest patient, daily supplementation of the 3 older patients was gradually reduced to 100 mg/Kg/day with a parallel decrease of Cr levels in blood and urine, but not in the brain where Cr and PCr/PDE remained stable or decreased only slightly (Fig. 4). The slight reduction in brain Cr levels observed at the dose of 100 mg/Kg/day in these patients seems to have no effect on their cognitive performances, which are stable in the same range (Fig. 3), Indeed, in the asymptomatic newborn, in which the initial dosage was lower (100 mg/Kg), a 65% Cr restoration in brain required a longer period (about 18 months) than in his affected relatives, in which about 90% of Cr replenishment was observed after 12 months of therapy [1, 10]. Considering that the blood brain barrier is more permeable in developing brain and Cr was well absorbed in all patients, as demonstrated by Cr concentrations in plasma, the differences in brain recovery time observed in the 4 patients could be ascribed to the different dosage regimens adopted. Low doses require more time to restore Cr in brain to the highest possible levels without affecting clinical improvement, thus avoiding peripheral compartment overloading and possible toxic effects. These findings seem to be confirmed by data of the other young AGAT-d patient [8], in which doses higher than 400 mg/Kg did not improve Cr/NAA ratio more than 70% of controls and increased the risk of renal damage. In our patients, reduction of doses to less than 100 mg/kg/day normalized Cr in blood and urine, with the best profile in P4, who is biochemically similar to his normal peers.

1H MRS acquisition of P4 at the age of 8 years showed a slight reduction of cerebral Cr as in P3 at the same age, but not in the female patients, despite their different dosage regimens.

In our patients, 100% recovery of cerebral Cr seems unfeasible, whatever treatment schedule was used; the highest Cr replenishment was found in P4 at about 4 years (98%). Although timing dosages for all reported AGAT-d patients are incomplete, we suppose that the principal factor accounting for this effect is the down regulation of the transporter protein at high Cr concentrations [5, 23]. However, in the mouse model of AGAT-d, total Cr brain accumulation continues over time up to the normal concentrations of wild-type mice [24], while in AGAT-d patients brain Cr always remains below control values. Partial recovery of cerebral Cr in patients could represent a missing amount deriving from cerebral AGAT-d synthesis. Endogenous Cr synthesis contributes to brain needs for a negligible part (no more than 20%) as demonstrated by mouse models of Cr transporter (CrT) deficit [24, 25], where it is reduced in both the cortex and hippocampus to about 18–20% of control groups. In addition, brain Cr in CrT deficient patients is absent or strongly reduced when assessed by 1H-MRS; thus supporting the hypothesis that endogenous synthesis does not compensate for loss of Cr deriving from periphery. These considerations might account for the partial brain Cr replenishment in AGAT-d patients which does not reach the same level of controls and remains stable below 90% even when Cr transport is forced with high doses. However, despite the partial restoration of brain Cr pool, the clinical status of these patients improve significantly during treatment and, in the one who was early treated, typical somatic and psychomotor development was observed.

Considering the best dosage regimen, we can hypothesize that oral Cr supplementation efficiently restores cerebral levels even at lower doses, although more slowly, after a short period at high doses (400 mg/kg/day) in older patients or at 100 mg/kg/day in younger ones. The small number of patients included in this study and their common genetic background, affecting both the clinical course and treatment response, do not permit us to suggest clinical guidelines suitable for all patients with AGAT-d. However, as in other rare inborn errors of metabolism, each patient constitutes his/her own control and for this reason the optimal dosage treatment response is not an unquestionable statement. Furthermore, clear evidence for the effect of Cr on clinical improvement in AGAT-d patients arises from the observation that patient status normalizes during treatment.

Despite the small number of patients, we conclude that 100% recovery of cerebral Cr seems unfeasible, whatever treatment schedule used, but there was no correlation between Cr levels and cognitive improvement. Cr treatment was well tolerated even at high doses, although side effects of weight gain and kidney stones were observed.

As indicated by the normal clinical development of P4, an early diagnosis is recommended for a timely treatment, suggesting the importance of NBS. Given the lack of a reliable biomarker in bloodspot for this inborn error of metabolism, it might be useful to consider other new options of treatment, such as untargeted metabolomics and/or genomic NBS. In the meantime, we should take in account AGAT-d in all patients presenting development delay, hypotonia, myopathy and language impairment in order to allow for an early diagnosis thus enabling treatment.