This study evaluates the correlation between the neurological progression of hepatic encephalopathy (HE) and amino acids and albumin levels in the CSF of rats with acute liver failure induced by subtotal hepatectomy. In this study, our group: i) proposes a scale for a reproducible approach to the neurological evaluation of rats with ALF; ii) correlates the increase of amino acids and albumin levels in CSF with animals’ neurological status; and iii) correlates the early (12 hours) increase of amino acid levels in CSF with a prognosis of death. Based upon these results, we hypothesize that the biochemical alterations observed in this study relate to BBB impairment in animals with HE, and, therefore, we propose that the increased levels of albumin and amino acids may be a potential biomarker for future medical evaluation and management of patients with ALF.
Acute liver failure is a life-threatening medical condition that frequently entails a significant brain impairment known as Hepatic Encephalopathy [1, 47, 48]. The reduction in hepatic workload leads to the accumulation of ammonia and other neurotoxins, which interfere with brain metabolism [49, 50]. In our experimental model, the hepatectomized animals present a significant increase of plasma levels in biomarkers of liver dysfunctionality: AST, ALT, Total Bilirubin, Direct Bilirubin, Prothrombin time and Ammonia, in comparison with Sham rats (Fig. 1) as observed in patients with ALF. In humans, Hepatic Encephalopathy symptoms range from mild alterations, such as ataxia, extremity tremor, mood changes, sleep disturbances and slurred speech up to severe neurological impairment, such as muscle rigidity, hyperreflexia, obtundation and even coma [51, 52]. Encephalopathy is clinically classified into degrees to better understand and monitor its severity [51, 53].
Animal models of liver disease have been studied for decades to understand the pathophysiology of HE, aiming to discover new therapeutic strategies for mitigating the impacts of this condition [54]. These experimental models reproduce the various stages of the disease for observation and classification purposes to identify the progression of neurological impairment subsequently. Mans and colleagues described a neurological scale for rats with HE induced by portocaval shunting and hepatic artery ligation [37]. More recently, Cauli and colleagues described another neurological grading method for rats with ALF induced by galactosamine [36]. These neurological scales, however, describe generic signs, such as mild ataxia, mild lethargy and poor gesture control to differentiate animals from stages 0, I and II. These criteria are susceptible to observational bias and may result in different interpretations [55, 37]. Our study proposes a modified neurological evaluation of HE, using simple and objective criteria with no additional costs or workload, which ensures better-defined criteria to sort subjects into neurological grades for HE in animal models. Our criteria are described in detail in the Methods – Neurological Evaluation section.
Elevated levels of free amino acids have been documented in both serum and CSF of animals with HE [37, 20, 56, 57]. These alterations seem to be consistent with elevated levels of glutamine and glutamate, which are also found in patients with HE [58, 59, 21, 22]. Similarly, our group previously described elevated levels of glutamate and glutamine in CSF both in a model of hyperammonemia and of acute HE [41, 35]. In this study, we observed a significant increase of all analyzed amino acids levels (except for alanine) in CSF in animals with ALF 24h after subtotal hepatectomy (Supplementary Table 1). These alterations were strongly correlated with the Grade of HE (Fig. 3). Animals with Grades I and II presented no alterations (or only mild elevations) in amino acid levels in CSF, while animals with Grades III and IV presented large increases of amino acid levels.
Additionally, albumin levels in CSF were correlated with the neurological grades of HE, increasing significantly in the late stages of the disease (Fig. 2). Albumin presents a high molecular weight and is unable to cross the BBB under normal conditions [60, 61]. Albumin levels in CSF only increase under pathological conditions [62]. Therefore, albumin levels in CSF are considered a marker of BBB permeability [63–65]. Hyperammonemia has been shown to affect BBB permeability and allows amino acids to enter into the brain [19, 24]. Previous studies in animal models of HE and ALF indicate mild alterations in BBB and do not demonstrate a clear breakdown of this brain system [66–68, 28, 27, 26]. Similarly, a study on patients with chronic liver disease and overt HE did not observe changes in BBB permeability, using positron emission tomography (PET) scans [69]. In the present study, however, we found elevated levels of almost all amino acids in CSF and an increase of albumin that was associated with the neurological deterioration due to HE. A diffuse penetration of various free amino acids into CSF is likely caused by a single factor, such as altered BBB permeability, rather than specific disruptions in the metabolic pathway of each amino acid caused by ammonia accumulation. This hypothesis is reinforced by the observed diffuse and homogenous increased penetration of Evans Blue staining into animals’ brain with HE, as shown in Fig. 2B.
The amino acid profile and albumin evaluation discussed above were performed 24h after subtotal hepatectomy – a result observed in previous studies carried out by our group [35, 39]. Following 24h of surgery, most of the rats present significant neurological impairment and are not yet in the final stages of HE, as the mean time of death is after 36 hours (Table 1). Another cohort of animals was selected for a second experiment to harvest CSF 12h after surgery. At this period, most animals presented HE Grade I, with few animals in Grade II (Table 1). We consider difficult to determine, via neurological parameters, which animals will improve their clinical condition and which will progress to coma and death. However, the evaluation of amino acid levels in CSF 12h after surgery allowed us to make a clear distinction between these two groups: the first group (Recovered group) presented a normal or mild elevation in amino acid levels and survived up to 72h (final period of the evaluation); while the second group (Deceased group) presented a marked elevation of amino acid levels in CSF (Fig. 4).
The glutamatergic system is believed to have an important role in ammonia detoxification in the brain, specifically via the production of glutamine [50, 35]. However, it is also believed that the accumulation of glutamine in the brain may cause dangerous effects, because of its involvement in the pathogenesis of brain edema and may act as a Trojan Horse by releasing ammonium into the mitochondria of the cells, thus disturbing brain bioenergetics [70–72]. Elevated levels of glutamine in serum and CSF have been associated with neurological decline as evidenced by EEG and with poor clinical outcomes in patients with HE [34, 23]. In this experiment, the evaluation of amino acid levels in CSF was of great value in differentiating healthy animals from those with HE, and animals with poor prognoses from those spontaneously recovered (Fig. 4). Considering this fact, elevated glutamine after 12 hours differentiated rats with HE from Sham-operated animals. Glutamine levels also allowed us to differentiate, with greater accuracy, the animals with favorable prognoses from those that would die before the end of the experiment (Fig. 4A). This was also true for methionine and phenylalanine (Figs. 4F and 4H, respectively). It is important to emphasize that the sample collection was performed 12 hours after surgery, a point at which it was not possible to predict the animals’ outcome evaluation by neurological grade of HE. These results correlate the increase of amino acids levels in CSF to an important outcome, such as mortality rate. We assume that if we could observe this same pattern of amino acid increased levels in CSF in humans with ALF, elevated glutamine levels could potentially be a useful biomarker for evaluating these patients.