Serotonin-related tryptophan in children with insulin-dependent diabetes

Introduction 

The brain serotonergic system is composed of a group of neurons located in the raphe nuclei in the mesencephalon and brainstem. Axons from these neurons innervate various important areas of the brain and spinal cord [1], [2]; the specific neurotransmitter is serotonin [5-hydroxytryptamine (5-HT)]. The existence and distribution of a serotonergic neuronal system in the brain of humans and rat fetuses and adults has been described [2], [3]. It has been established that serotonin plays roles of a neurotransmitter and a neuromodulator activating a large family of G proteins coupled to metabotropic-specific receptors in target neurons of various important areas of the brain [4], [5]. In addition, much evidence supports the participation of serotonin as a trophic and differentiation signal during nervous system development [5], [6], [7]. Furthermore it has been implicated in several important functions in the adult brain [8], [9]. On the other hand, brain 5-HT has been found to be involved in a number of neurologic and psychopathologic states [10], [11], [12], [13]. Because of the relevance of 5-HT function in the young and adult brain its early alteration could be significant for the normal development and function of related neuronal paths.

An essential amino acid, l-tryptophan (l-Trp), a protein constituent of a normal diet, is the precursor nutrient for the synthesis of brain serotonin. There are two known fractions of plasma l-Trp, one bound to albumin and one free (FFT) [14]. FFT passes through the blood-brain barrier to the brain, where it is taken up by serotonergic neurons and hydroxylated by the action of tryptophan-5-hydroxylase (TrpOH). Then, 5-hydroxytryptophan is formed, decarboxylated, and transformed into 5-HT [15]. There are several possible mechanisms proposed for the regulation of the amount of plasma l-Trp passing to the brain. First, a specific transport system has been postulated [16]. Second, the amount of FFT available to be transported to the brain would depend on its binding to albumin [14]. This mechanism may compete at the blood-brain barrier level, with a higher affinity of the carrier system for l-Trp at the brain capillary wall, which would strip the amino acid from albumin and increase its transport toward the brain [17]. Third, neutral amino acids (NAA), phenylalanine, valine, leucine, tyrosine, and isoleucine (Phe, Val, Leu, Tyr, Ile) appear to share the same carrier at the blood-brain barrier and would also compete with l-Trp for transport to the brain [18].

There is experimental evidence in rats with diabetes mellitus that points to a decrease in brain serotonin coincident with a lowering of plasma FFT and of brain l-Trp and in the activity of TrpOH, which supports a diminished activity of the serotonergic system [19], [20], [21], [22], [23], [24], [25].

With these points in mind, we proposed the hypothesis that school children with insulin-dependent diabetes mellitus may have a modified relation between FFT and NAA in their plasma, compared with normal children of the same age. Because quantification of this fraction is possible in human plasma, it can be useful as a peripheral measurable indicator of brain 5-HT synthesis, as we have demonstrated in early malnourished patients [26], [27]. To test this hypothesis a comparative study was performed with two groups of school children, one group including children with insulin-dependent diabetes mellitus, and the other group consisting of normal children.

Patients and methods 

In the present study, 34 children, 6.83-10.49 years of age, were selected from the Endocrinology Service of the Pediatric Hospital, XXI Century, National Medical Center, Mexican Institute of Social Security, Mexico City, Mexico. Two groups were formed. The first group included 22 children with insulin-dependent diabetes mellitus, according to the National Diabetes Group [28] criteria, with a body mass index (weight/length2) normal for their age and without other underlying diseases. The second group was made up of 12 normal children within the same age range who served as control subjects. None of the children with diabetes was in the remission period or under treatment with salicylates, chlorofibrates, benzoates, or indomethacin nor did they have acute complications of the disease, such as electrolyte imbalance, diabetic ketoacidosis, or infection.

All children were fed a normal diet of 55 kcal/kg/day (protein 30%, carbohydrates 55%, lipids 15%). Additionally, patients with diabetes were treated with a mixture of fast and intermediate-action insulin, 0.8-1 U/kg/day. Three mL of blood were collected by venopuncture in borosilicate tubes containing 450 μL of ACD solution, which consisted of 3.6 mg sodium citrate, 9.9 mg citric acid, 11 mg dextrose, buffered with 50 mmol Tris acetate, pH 7.40 between 07:30 and 08:30 am and 12 hours after the last feeding. The tubes containing the blood samples were cooled immediately (0-4°C) on ice and centrifuged at 500 g in a refrigerated centrifuge to obtain the plasma sample. Aliquots were taken for the following biochemical assays: 100 μL for FFT, 20 μL for total l-Trp (the difference between total and FFT was considered as the albumin-bound fraction), 200 μL for NAA, 25 μL for albumin, 50 μL for free fatty acids, 20 μL for glucose, and 50 μL for glycosylated hemoglobin.

Results 

Children with insulin-dependent diabetes mellitus had an evolution of 4.4 ± 2.7 years of age when the study was performed. There were no differences in anthropometric data when compared with control children (Table 1). Glycemia and glycosylated hemoglobin were significantly elevated (P < 0.001) in patients with diabetes. Furthermore, these children had an increase in free fatty acids (P < 0.01) (Table 2) and NAA in plasma, when compared with normal children (P < 0.05) (Table 3). Note that the plasma albumin of both groups of children was similar (Table 2).

Table 1. Clinical data of school children with insulin-dependent diabetes mellitus and control subjects

	Data	Insulin-Dependent Diabetes Mellitus	Control Subjects
	Age (yr)	8.66 ± 1.83	8.25 ± 1.05
	Sex
	Males	10	8
	Females	12	4
	Body weight (kg)	30.82 ± 6.75	27.11 ± 5.46
	Length (m)	1.3 ± 0.1	1.28 ± 0.03
	Body mass index	17.35 ± 2.0	17.13 ± 2.18
	Time of evolution (yr)	4.41 ± 2.7	–

Each point represents the mean value ± S.D. Differences were determined by Student t test.

Table 2. Biochemical data in plasma of school children with insulin-dependent diabetes mellitus and control subjects

	Data	Insulin-Dependent Diabetes Mellitus	ControlSubjects
	Glucose (mg/dL)	209.14 ± 10.66*	90.55 ± 15.61
	Glycosylated hemoglobin (%)	9.23 ± 3.37†	4.65 ± 0.54
	Albumin (g/dL)	4.63 ± 0.1	5.00 ± 0.25
	Free fatty acids (mmol/mL)	1.75 ± 0.16†	0.55 ± 0.02

Each point represents the mean value ± S.D. of 34 determinations from 22 insulin-dependent diabetes mellitus patients and 12 control subjects. All determinations were performed in duplicate. Differences were determined by Student t test.

* P < 0.001. † P < 0.01.

Table 3. Plasma concentration of neutral amino acids in school children with insulin-dependent diabetes mellitus and control subjects

	Amino Acids	Insulin-Dependent Diabetes Mellitus	Control Subjects
	Valine	112.3 ± 12.07*	80.65 ± 5.48
	Isoleucine	78.86 ± 11.08*	46.43 ± 4.7
	Leucine	78.86 ± 11.08*	46.43 ± 4.7
	Phenylalanine	49.86 ± 6.63*	33.75 ± 3.21
	Tyrosine	66.76 ± 12.03*	30.69 ± 6.37

Each point represents the mean value ± S.D. (μmol) of 34 determinations from 22 insulin-dependent diabetes mellitus patients and 12 control subjects. All determinations were performed in duplicate sample. Differences were determined by Student t test.

* P < 0.01.

One of the most important parameters measured in these children was the different fractions of l-Trp in plasma. School children with diabetes demonstrated a significant decrease of FFT, bound to albumin and total l-Trp, as compared with normal children (P < 0.001). Additionally, values of FFT/total l-Trp and FFT/NAA ratios were significantly lower in the insulin-dependent diabetes mellitus group (P < 0.001) (Table 4).

Table 4. Plasma concentration of l-trypytophan in school children with insulin-dependent diabetes mellitus and control subjects

	l-Tryptophan	Insulin-Dependent Diabetes Mellitus	ControlSubjects
	Free fraction	3.49 ± 0.92*	10.49 ± 0.56
	Albumin bound	36.46 ± 5.52*	47.24 ± 2.96
	total	39.61 ± 5.05*	57.75 ± 2.92
	Free fraction/total ratio	0.088 ± 0.06*	0.181 ± 0.012
	Free fraction/neutral amino acid ratio	0.154 ± 0.067*	0.365 ± 0.105

Each point represents the mean value ± S.D. (μmol) of 34 determinations from 22 insulin-dependent diabetes mellitus patients and 12 controls. All determinations were performed in duplicate. Differences were determined by Student t test.

* P < 0.001.

Discussion 

Results of biochemical analyses confirm the hypothesis that school children with insulin-dependent diabetes mellitus have a decrease of FFT in plasma. The observed decrease in FFT in the diabetic patients, can not be explained by the increase in free fatty acis that would tend to favor its increase, because free fatty acids compete with l-Trp for binding to albumin [14], [35]. Pardridge [17] pointed out the importance of the final concentration of FFT at the blood-brain barrier capillary level, which could be higher than the fraction measured in vitro, because of possible higher affinity of the brain capillary carrier for the amino acid, as compared with that of albumin. Albumin concentration did not change, thus the binding mechanism of the amino acid to albumin itself may be also modified, as observed in malnourished children [26], [27]. The concomitant increase of NAA in plasma would also play an interfering role for l-Trp transport to the brain [36], [37], [38]. Therefore the decreased availability of FFT at the blood-brain barrier would decrease its transport to the brain, so that the resulting balance would be a reduction of serotonin synthesis.

A decrease in FFT can also be explained by a deviation of l-Trp to alternative metabolic pathways, such as those of kynurenic and nicotinic acid [39], which could mask the expected increase, with a final decrease at the blood-brain barrier level. Furthermore, according to Sadler et al. [40], in the diabetic state, it may explained by an increased catabolism of l-Trp, after stimulation of liver tryptophan oxygenase activity. NAA in plasma of school children with insulin-dependent diabetes mellitus, demonstrated a consistent change as compared with control subjects, rendering a decrease of FFT and a lower FFT/NAA ratio. Altogether these metabolic changes may contribute to a lower synthesis of the brain neurotransmitter [39], [40], [41].

In experimentally induced diabetic rats, brain serotonin synthesis is decreased, coincident with a smaller concentration of FFT in plasma [20], [24], [25]. The present results demonstrate that this condition is also present in human school children with insulin-dependent diabetes mellitus, with FFT/total l-Trp and FFT/NAA ratios also low. To our knowledge, this is likely the first observation reported of an alteration of the precursor amino acid, together with other metabolites related to the synthesis of brain serotonin in human patients with diabetes.

In the present study, we obtained additional information concerning the mechanisms of damage to human brain caused by the effects of insulin-dependent diabetes mellitus on a specific neuronal system. In addition, we have pointed out its possible negative influence that could alter brain function permanently. These results may have clinical relevance, because brain serotonin is known to play an important role in the pathophysiology of various neuropsychiatric disorders, such as depression, anxiety, panic, obsessive-compulsive disorder, social phobia, schizophrenia, anorexia nervosa, and other related problems, such as aggression, addictions, impulse control, and attention deficit [10], [11], [12], [13], [42], [43]. In our laboratory, we are currently studying the possible functional correlates of presumed 5-HT dysfunction in the brain of diabetic children.