Diabetes insipidus is characterized by the excretion of abnormally large volumes of dilute urine under conditions of ad libitum fluid intake. In rare cases, it is caused by mutations in genes encoding proteins essential for proper regulation of the outflow of water. During the last two decades, the knowledge regarding the genes involved in disease development has advanced tremendously. Our ability to identify mutations in the patients and predict their pathogenic effect represents a major advance in clinical care. It eliminates the cost and inconvenience of repeated clinical monitoring of infants who are not at risk and allows full attention to be given to early detection and treatment in those who are virtually certain to manifest the disease. It further enables verification of the diagnosis and genetic counseling. This chapter aims to provide a brief overview of clinical and genetic aspects of different types of diabetes insipidus, the differential diagnoses, and information about the indications and practicalities of genetic testing in diabetes insipidus available today.
Keywordsdiabetes insipidus, polyuria, polydipsia, neurohypophyseal, nephrogenic, AVP, WFS1, PCSK1, AVPR2, AQP2
Disorders of water balance are common, and practicing physicians are often met with complaints of increased urination and thirst. Only a minority of such patients, however, suffers from diabetes insipidus and in even fewer the symptoms are caused by genetic defects in any one of the genes being essential in ensuring proper water homeostasis of the human body. Nevertheless, such patients should be identified and subjected to proper clinical and genetic testing in order to secure correct diagnosis and optimal treatment. The clinical differential diagnosis of diabetes insipidus can be challenging, and there are multiple examples of misdiagnosis, especially when the disease presents in a partial form. Although familial forms of diabetes insipidus were recognized more than 150 years ago, the genetic causes have only been revealed in recent decades. At present, genetic testing represents not only an important tool in the differential diagnosis of diabetes insipidus, but also a major advance in clinical care, as it has to some extent eliminated the problem of trying to identify at birth which offspring are likely to develop the disease.
The term “diabetes insipidus” is derived from the Greek “diabainein,” meaning to pass through, and from Latin “insipidus,” meaning tasteless. A distinction between diabetes associated with sweet tasting urine (diabetes mellitus) and diabetes associated with insipid (tasteless) urine (diabetes insipidus) seems to have been firmly established in 1831, but as early as in 1790, a case of diabetes in which “the urine was perfectly insipid” was recorded in the literature and noted to be “very rare; and that the other (read: diabetes mellitus) is by much the more common.”
The prevalence of diabetes insipidus is yet not clearly established but has been estimated to be approximately 1:25,000 and with an annual incidence of approximately 0.01%. Of these, less than 10% are inherited.
Diabetes insipidus is characterized by the excretion of abnormally large volumes of dilute urine under conditions of ad libitum fluid intake. Diabetes insipidus is defined clinically by the following two criteria:
24-h urine volume exceeding 50 mL/kg body weight (in children 75–100 mL/kg body weight, due to higher water content in their food)
Urinary osmolality < 300 mosmol/kg
The polyuria is distinguishable from the osmotic diuresis of uncontrolled diabetes mellitus or other forms of solute diuresis by the absence of glucosuria and a normal rate of solute excretion (10–20 mosmol/kg per day).
This chapter aims to provide a brief overview of clinical and genetic aspects of different types of diabetes insipidus and the differential diagnosis and information about the indications and practicalities of genetic testing in diabetes insipidus available today.
Types of diabetes insipidus
Diabetes insipidus can be divided into four different types that are caused by any one of four fundamentally different defects ( Fig. 5.1 ): 1. pituitary , central , neurogenic , or neurohypophyseal diabetes insipidus, the most common type, results from a deficiency in the production of the antidiuretic hormone arginine vasopressin (AVP); 2. renal or nephrogenic diabetes insipidus is caused by renal insensitivity to the antidiuretic effects of AVP, for example, due to impairment of the renal vasopressin V2 receptor or aquaporin-2 water channel; 3. primary polydipsia is due to suppression of AVP secretion as a result of excessive fluid intake. Depending on whether the excessive fluid intake is due to abnormal thirst or due to a psychological disorder, primary polydipsia is subdivided into, respectively, dipsogenic diabetes insipidus psychogenic diabetes insipidus, and 4. gestational diabetes insipidus, which is primarily due to increased metabolism of AVP by circulating vassopressinase produced by the placenta in the pregnant woman but may also involve renal resistance and/or subclinical deficiency in AVP production.
Complete diabetes insipidus is defined by persistently low urine osmolality (<300 mosmol/kg) during a fluid deprivation test providing plasma osmolality rises above 295 mosmol/kg. Partial diabetes insipidus is defined by a subnormal increase in urine osmolality (300–600 mosmol/kg) during a fluid deprivation test with the same rise in plasma osmolality.
Familial types of diabetes insipidus
Neurohypophyseal and nephrogenic diabetes insipidus are in rare cases inherited and thus referred to as familial neurohypophyseal diabetes insipidus (FNDI) and congenital nephrogenic diabetes insipidus (CNDI or NDI). Based upon further clinical characteristics and inheritance patterns, FNDI can be subdivided into six different forms and NDI into three ( Table 5.1 ). The underlying genetic defects have been identified in the majority of these different forms, and in FNDI they include mutations in either the AVP gene (Entrez Gene: 551) encoding the AVP prohormone (vasopressin-neurophysin 2-copeptin), the WFS1 gene (Entrez Gene: 7466) encoding wolframin or in the PCSK1 gene (Entrez Gene: 5122) encoding proprotein convertase 1/3 (neuroendocrine convertase 1). In NDI, they include mutation in either the AVPR2 gene (Entrez Gene: 554) encoding the vasopressin V2 receptor or the AQP2 gene (Entrez Gene: 359) encoding aquaporin-2. In the following sections, we will summarize the main clinical characteristics and genetic aspects of the four most common forms of inherited diabetes insipidus: 1. autosomal dominant FNDI , 2. autosomal recessive FNDI , type c (in Wolfram syndrome) , 3. X-linked recessive NDI , and 4. autosomal dominant/recessive NDI ( Table 5.1 ).
|Disease||Inheritance pattern||Chromosomal location||Affected gene||OMIM||Number of kindreds|
|Autosomal recessive, type a||20p13||AVP||125700||2|
|Autosomal recessive, type b||20p13||AVP||NR||1|
|Autosomal recessive, type c||4p16.1||WFS1||222300||>175 *|
|Autosomal recessive, type d||5q15||PCSK1||NR||8 **|
Autosomal Dominant FNDI
The familial occurrence of severe polyuria and polydipsia (up to 28 L/24 h), segregating in an autosomal dominant pattern, and responding readily to exogenous dDAVP (autosomal dominant FNDI) (OMIM: 125700) shows several intriguing features that separate it from other familial forms of diabetes insipidus ( Table 5.2 ). Usually, affected family members show no signs of disturbed water balance at birth and during early infancy but develop progressive symptoms of excessive drinking and polyuria at some point during infancy or early childhood. Even at this stage, the deficiency of AVP, although marked, is often incomplete, and can be stimulated by hypertonic dehydration, leading to concentration of the urine. This may lead to delayed recognition of the condition and even misdiagnosis. In the few cases in which it has been studied by repetitive fluid-deprivation tests, AVP production is normal before the onset of FNDI but diminishes progressively during early childhood. Once fully developed, the diabetes insipidus with severe thirst, polydipsia, and polyuria (8–20 L/day) is similar to the diabetes insipidus in other complete forms. In a few patients, yet, the AVP deficiency remains partial for decades, even though it is complete in other affected members of the same family with the same disease causing mutation, implying that other factors, genetic and/or environmental, are in play to modulate the penetrance of the mutations. That could, for example, be differences between families (and individuals) in their ability/awareness to ensure early fluid replacement in affected children, which would be expected to delay the time of onset and slow down disease progression. In some middle-aged male patients, diabetes insipidus symptoms decrease markedly without treatment and with preserved AVP deficiency and normal glomerular filtration. The mechanism of these remissions is currently unexplained. Consistent with potential loss of AVP-producing magnocellular neurons, the hyperintense MRI (magnetic resonance imaging) signal normally emitted by the posterior pituitary is absent or very small in patients with autosomal dominant FNDI, at least by the time AVP deficiency becomes clinically recognizable. Since the same lack of signal is characteristic in patients with NDI, the usefulness of this investigation in the differential diagnosis is questionable.
|Autosomal dominant||X-linked recessive||Autosomal dominant/recessive|
|Male:female ratio||1:1||Males only *||1:1|
|Debut of symptoms||6 months to 6 years||From birth **||From birth †|
|Plasma AVP during thirst||Low/undetectable||High||High|
|Decrease of symptoms at middle age||Yes, in some cases||NR||NR|
|Antidiuretic response to dDAVP||>50% increase in U-osm||<50% increase in U-osm||<50% increase in U-osm|
|Extrarenal response to dDAVP (e.g., factor VIII)||Normal||Reduced||Normal|
|Posterior pituitary bright signal on MRI||Lacking||Lacking||NR|
The clinical characteristics of the rare autosomal recessive forms of FNDI (type a and type b) and autosomal recessive FNDI in congenital malabsorptive diarrhea (type d) ( Table 5.1 ) differ in many aspects from those of autosomal dominant FNDI, for example, regarding age of onset, plasma levels of AVP during fluid deprivation, inter-/intrafamily variation, and cooccurrence with a complex picture of other symptoms. For a thorough discussion of some of these rare conditions we refer to Refs .
Nonsyndromic FNDI has been reported in more than 100 kindreds worldwide, and in the majority of these, the disease is caused by mutations in the AVP gene ( Table 5.1 ). With only a few well-documented exceptions, FNDI is transmitted by autosomal dominant inheritance and appears to be largely, if not completely, penetrant with age. Based upon the initial report in 1945 by Forssman it has been listed in databases (OMIM: 304900) that an X-linked recessive form of FNDI exists. However, a reinvestigation, including genetic and clinical examinations on descendants of the original patients, has revealed that the original patients more likely had a partial NDI phenotype, and consistent with this the reinvestigated patients actually carried a mutation in the AVPR2 gene (g.310C > T, p.Arg104Cys). On the other hand, in another report of X-linked recessive inheritance of FNDI ( Table 5.1 ), the clinical phenotype in one family is clearly consistent with that of neurohypophyseal diabetes insipidus, and although the disease seems to be linked to the same chromosomal location as the AVPR2 gene (Xq28), mutations have been found neither in this gene nor in the AVP gene. For a more thorough discussion of this rare condition, we refer to Ref. .
Until now, FNDI has been associated with >70 different mutations in the AVP gene. All but one of these mutations (g.1919 + 1delG) disturbs the coding region of the gene. The type and location of the mutations differ widely and affect amino acid residues ranging from the N-terminal residue of the signal peptide (the start codon), through the AVP moiety and into the C-terminal part of the neurophysin 2 domain of the AVP preprohormone. Of note, however, FNDI has not been attributed to any mutation affecting solely copeptin, the glycosylated peptide encoded by exon 3. A copeptin mutation (p.Ala159Thr) has been identified in one family in which autosomal dominant FNDI was segregating. It did not, however, segregate with the disease and all affected family members were heterozygous for another mutation (p.Gly96Asp) previously reported to cause FNDI. This strongly suggests that the p.Ala159Thr mutation had no pathological effect.
There is no real prevalent mutation in the AVP gene in autosomal dominant FNDI but one mutation is nonetheless more frequent than all others, namely, the g.279G > A substitution (predicting an p.Ala19Thr amino acid substitution), which has been identified in 11 kindreds with no known relationship to each other.
The genetic basis of autosomal dominant FNDI remains unknown in one reported kindred. In this Chinese family, the disease showed linkage to a 7-cM interval on chromosome 20p13 containing the AVP gene; however, unexpectedly, no mutations could be found in the coding regions, the promoter, or the introns of the AVP gene. In addition, the coding regions of the nearby UBCE7IP5 gene and the AQP2 gene on chromosome 12 were analyzed, but no mutations were detected. Thus, the cause of FNDI in this specific family remains unexplained.
The diversity of the mutations identified in the AVP gene in autosomal dominant FNDI could indicate diverse pathogenic mechanisms. However, the fact that different disease alleles result in rather uniform clinical presentations calls for a single unifying explanation. One such explanation suggests that mutations in the AVP gene exert a dominant-negative effect by leading to the production of a mutant AVP prohormone that fails to fold and/or dimerize properly in the endoplasmatic reticulum (ER) and, as a consequence, is retained by the ER protein quality control resulting in cytotoxic accumulation of protein in the magnocellular neurons that produce AVP (i.e., a misfolding-neurotoxicity hypothesis).
Autosomal Recessive FNDI, Type C
Autosomal recessive inheritance of FNDI has been described and is by far most common in conjunction with Wolfram syndrome (OMIM: 222300) ( Table 5.1 ). The syndrome is a rare, multifaceted, and progressive neurodegenerative disease also referred to as DIDMOAD, abbreviating its symptom spectrum, including d iabetes i nsipidus, early-onset, insulin-dependent d iabetes m ellitus, progressive o ptic a trophy, and sensorineural d eafness, combined with many other neurological abnormalities. It has been estimated that approximately 75% of patients with Wolfram syndrome present with partial neurohypophyseal diabetes insipidus at an average age of 14 years (range: three months to 40 years of age). This is about the same time as they lose their hearing, but after onset of diabetes mellitus and loss of vision. The polyuria in Wolfram syndrome is due to a partial or severe deficiency in AVP production, and seems to be associated with posterior pituitary degeneration and impaired processing of the AVP prohormone. In general, the patients respond well to dDAVP administration. Accordingly, the diabetes insipidus observed in Wolfram syndrome seems to be clinically distinguishable from autosomal dominant FNDI only by its mode of inheritance, a tendency toward higher age of onset, and most obvious by its cooccurrence with a complex picture of other symptoms. Besides, in clear contrast to the rather mild course of autosomal dominant FNDI, patients with Wolfram syndrome usually die from central respiratory failure as a result of brainstem atrophy in the third or fourth decade of life.
Patients with the Wolfram syndrome are often offspring from consanguineous marriages between unaffected parents, and they have affected siblings, strongly supporting a recessive pattern of inheritance. Most commonly, patients are homozygous or compound heterozygous for mutations in the WFS1 gene located on chromosome 4 (4p16.1). Infrequently, mutations in the CISD2 gene can also result in Wolfram syndrome; however in these cases diabetes insipidus seems to be absent.
The protein product Wolframin, encoded by WFS1 , has been shown to play a crucial role in the negative regulation of a feedback loop of the ER stress–signaling network, induced by the accumulation of misfolded and unfolded proteins in the organelle. More specifically, it seems to prevent secretory cells, such as pancreatic β-cells, from death caused by dysregulation of this signaling pathway. It could be that Wolframin plays a similar role in the AVP producing magnocellular neurons. These cells may experience ER stress as an effect of the large biosynthetic load placed on the ER by the exceptional high demands for AVP prohormone production under physiological conditions that stimulate the neurons (elevated osmotic pressure), exactly as pancreatic β-cells during postprandial stimulation of proinsulin biosynthesis. Loss of Wolframin function under such conditions would lead to unregulated ER stress signaling and probably neuronal cell death, exactly as has been shown in pancreatic β-cells. This mechanism links directly to the ER stress, possibly elicited by the accumulation of misfolded mutant AVP prohormone in the magnocellular neurons of patients with autosomal dominant FNDI.
X-Linked Recessive NDI
NDI is characterized by the same symptoms of diabetes insipidus as seen in FNDI but shows renal insensitivity to the antidiuretic effect of endogenously produced AVP as well as to dDAVP treatment ( Table 5.2 ). As mentioned earlier, it is inherited either in an X-linked recessive pattern (in approximately 90% of NDI cases) (OMIM: 304800) and caused by mutations in the AVPR2 gene or in an autosomal dominant or recessive pattern (in approximately 10% of NDI cases) (OMIM: 125800) and caused by mutations in the AQP2 gene ( Table 5.2 ). Inherited NDI additionally exists in complex forms characterized by loss of water and ions, for example, Bartter syndrome (OMIM: 601678, 241200, 607364, and 602522). These rare variant forms of NDI are caused by mutation in genes ( KCNJ1 , SLC12A1 , CLCNKB , CLCNKB , and CLCNKA in combination, or BSND ) encoding other renal tubular transporters, resulting in a disturbance of the inner medullary concentration mechanism and thereby polyuria and polydipsia together with characteristic electrolyte disturbances. The electrolyte disturbances, however, clearly distinguish these conditions from genuine diabetes insipidus.
The clinical characteristics of NDI attributable to mutations in the AVPR2 or AQP2 gene, regardless of the mode of inheritance or specific mutation involved, include hypernatremia, hyperthermia, mental retardation, and repeated episodes of dehydration if patients cannot obtain enough water. In contrast to autosomal dominant FNDI, the urinary concentration defect is present at birth, and newborns and infants are especially prone to dehydration ( Table 5.2 ). Furthermore, the clinical picture of dehydration is often difficult to interpret at this age (usually manifested as failure to thrive), and therefore severe cases of prolonged dehydration are seen. Mental retardation was prevalent in 70–90% of the patients reported in the original studies of NDI and was thought to be part of the disease. However, early recognition of NDI based on genetic screening of at-risk children and early treatment of the disease permit such children to have normal physical and mental development and suggest that the mental retardation reported in the original studies probably resulted from repeated episodes of severe dehydration rather than from pathologies directly caused by the genetic defect.
The X-linked recessive and most frequent form of familial NDI is caused by loss-of-function mutations in the AVPR2 gene, encoding the renal vasopressin V2 receptor. Currently, at least 214 putative disease-causing mutations in the AVPR2 gene have been identified in more than 300 families. Usually, heterozygous females are asymptomatic but may in some cases have variable degrees of polyuria and polydipsia because of skewed X-chromosome inactivation. Recent reports of patients having mutations in the AVPR2 gene and initially suspected of having FNDI due to in some cases an impressive antidiuretic response to exogenous AVP have emphasized the importance of genetic testing and have added to the complexity of clinical differential diagnosis in diabetes insipidus.
The molecular mechanism underlying the renal AVP insensitivity in NDI differs among disease alleles. Hence, AVPR2 mutations have been divided into five different classes according to their pathogenic effect at the cellular levels, exactly as the system used for classification of mutant low-density lipoprotein (LDL) receptors. Most mutations in the AVPR2 gene (>50%) result in vasopressin V2 receptors that are trapped inside the cell due to impaired intracellular trafficking, and they are consequently unable to reach the plasma membrane and bind the ligand (AVP) (Type 2 mutations). Other mutations lead to an unstable mRNA, and thereby no protein product (Type 3 mutations) or mutant vasopressin V2 receptors that reach the cell surface, but either cannot bind AVP (the ligand) efficiently (Type 1 mutations) or properly trigger an increase in intracellular cyclic AMP (cAMP) production (Type 4 mutations).
Treatment of patients with NDI with dDAVP is usually not effective, although females having symptoms due to skewed X-chromosome inactivation and patients with partial NDI can be alleviated from their diabetes insipidus by using high doses, possibly because such patients retain some functional vasopressin V2 receptors. Low sodium diets to reduce the solute load to the kidneys and treatment with diuretic thiazide eventually in combination with cyclooxygenase inhibitor (indomethacin) or hydrochlorothiazide in combination with amiloride represents treatment regimens that to some extent can relieve symptoms of NDI. Novel strategies suggested for the treatment of NDI are: vasopressin V2 receptor antagonists and agonists, pharmacological chaperones, nonpeptide agonists, cyclic guanosine monophosphate pathway activation, cAMP pathway activation, statins, prostaglandins, and molecular chaperones. For a detailed discussion of these approaches, we refer to Ref. .
Autosomal Dominant/Recessive NDI
As mentioned earlier, the clinical characteristics of NDI attributable to mutations in the AVPR2 or AQP2 gene, regardless of the mode of inheritance or specific mutation involved, are similar ( Table 5.2 ). Autosomal dominant or autosomal recessive modes of inheritance occur in approximately 10% of the families with inherited NDI, with the autosomal recessive form being the most common among the two (90% vs. 10%). Typically, these families have mutations in the AQP2 gene, encoding the aquaporin-2 water channel. Currently, at least 51 putative disease-causing mutations in the AQP2 gene have been identified, mostly in children of consanguineous parents.
In vitro studies show that different mutations in the AQP2 gene have different pathological effects at the cellular level. Misfolding of mutant aquaporin-2 proteins and subsequent degradation in the ER is likely the major mechanism underlying autosomal recessive NDI, whereas mutations causing autosomal dominant NDI generally affect the carboxy-terminus of aquaporin-2, causing misrouting of both mutant and wild-type aquaporin-2 proteins. Thus, these mutations probably act by a dominant negative mechanism by preventing the normal aquaporin-2 protein from reaching the apical surface of the collecting duct epithelial cells. Some of these mutations result in a partial NDI phenotype, which has also been observed in patients with mutations in the AVPR2 gene. For a detailed discussion of partial NDI in patients with AVPR2 or AQP2 mutations, we refer to Ref. .