The molecular biology of human hereditary central diabetes insipidus

D.F. Swaab, M.A. Hofman, M. Mirmiran, R. Ravid and F.W. van Leeuwen (Eds.) Progress in Brain Research, Vol. 93 0 1992 Elsevier Science Publishers B.V. All rights reserved.

295

CHA$TER 21

The molecular biology of human hereditary central diabetes insipidus David R. Repaske and John A. Phillips, 111' Division of Pediatric Endocrinology, Department of Pediatrics, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7220, U.S.A.; and I Division of Genetics, Department of Pediatrics, Vanderbilt University, Nashville, TN 37232, U.S.A.

Introduction Diabetes insipidus (DI) is a disease characterized by excessive excretion of water in the urine. In this disorder hypothalamic sensing of plasma osmolality is uncoupled from renal mechanisms for conserving water. The hypothalamus normally communicates fluid status to the kidney via regulation of release of a posterior pituitary hormone, arginine vasopressin (AVP), otherwise known as antidiuretic hormone. AVP is a nine amino acid peptide hormone with one intramolecular disulfide bond. It is synthesized in magnocellular neurons of the supraoptic and paraventricular nuclei of the hypothalamus, and axonally transported in neurosecretory granules with its carrier protein, neurophysin 11, and a glycoprotein to nerve terminals in the posterior pituitary (Brownstein, 1983). AVP is normally secreted into the circulation as the free peptide in response to plasma hyperosmolality sensed directly in the hypothalamus or in response to peripheral stimuli communicated to the hypothalamus by afferent neurons (Sawchenko and Swanson, 1981). Decreased extracellular fluid volume is sensed primarily by atrial stretch receptors, and hypotension by aortic baroreceptors. There are also humoral influences that potentiate AVP secretion, including angiotensin I1 (Iovino and Steardo, 1984), 0-adrenergic hormones (Schrier et al., 1979), and possibly others (Zerbe and Robertson, 1987). AVP acts on epithelial cells of the collecting

tubules of the kidney to increase their water permeability, allowing reabsorption of water from the renal ultrafiltrate. This action is mediated by a membrane-bound AVP receptor that stimulates an intracellular adenylate cyclase to increase the intracellular cyclic AMP (CAMP)concentration (Jard et al., 1975). CAMP-dependent kinases then catalyze phosphorylations that promote insertion of preformed intracellular stores of water channels into the cell membrane. The increase in the water permeability of the collecting tubule epithelium allows the resorption of water from the dilute ultrafiltrate to the hypertonic renal medullary interstitium (Dousa et al., 1977; Brown and Orci, 1983). In addition, AVP, also by a CAMP-dependent mechanism, acts in the medullary thick ascending limb of Henle to enhance Na+ and C1- transport from the ultrafiltrate to the renal medulla, thereby further increasing the medullary hypertonicity and the ability of the kidney to resorb water (Herbert and Andreoli, 1984). DI can result from one of two distinct types of defects. It can be central, that is, due to insufficient release of AVP from the posterior pituitary, or nephrogenic, that is, due to inadequate response of the kidney to circulating AVP. Both types have multiple etiologies, but they can usually be distinguished by assay of circulating AVP when the plasma has a relatively high osmolality and/or by testing the effect of exogenous AVP or AVP analog on the ability of the kidney to produce a concentrated urine. In central

296

DI, circulating AVP is inappropriately low, and the kidney will concentrate urine in response to exogenous AVP. In nephrogenic DI, the circulating AVP concentration is appropriate for plasma osmolality, and therefore additional exogenous AVP has no additional effect on the urine. Nephrogenic DI can be inherited as an X-linked recessive disorder (Bode and Crawford, 1969), but is usually sporadic or drug-induced (Reeves and Andreoli, 1989). Clinically, the majority of cases of central DI result from damage to the hypothalamus, pituitary stalk, or posterior pituitary. Most cases are secondary to head trauma, surgical procedures, or inflammatory or neoplastic disease. A small minority of cases of human central DI are familial. X-linked inheritance has been described in a limited number of families (Forssman, 1955; Green et al., 1967). Usually, inheritance follows an autosomal dominant pattern (Pender and Fraser, 1953; Levinger and Escamilla, 1955; Moehlig and Schultz, 1955; Martin, 1959; Braverman et al., 1965; Meinders and Bijlsma, 1970; Kaplowitz et al., 1982; Blackett et al., 1983; Toth et al., 1984; Pedersen et al., 1985). Patients with autosomal dominant central DI typically have polyuria and polydipsia with variable clinical severity in different affected members of a single family and even variable severity in an individual over time (Pender and Fraser, 1953; Martin, 1959; Green et al., 1967; Kaplowitz et al., 1982). Onset of clinical symptoms occurs from infancy to several years of age, and some patients report amelioration of symptoms in the third to fifth decades. Biochemical, histological and genetic investigations

The first measurements of plasma AVP concentration in human autosomal dominant central DI were reported by Kaplowitz et al. (1982) who found marked deficiencies in two affected brothers. The severely affected brother had no detectable circulating AVP, even after fluid restriction that raised his plasma osmolality well above normal to 306 mOsm/kg. The less severely affected brother was able to produce a small amount of AVP in both water deprivation and hypertonic saline infusion

tests. This paper documented a relationship between the severity of disease in individual family members and the degree of vasopressin deficiency in these individuals. A limited number of autopsy studies (Braverman et al., 1965; Green et al., 1967; Nagai et al., 1984; Bergeron et al., 1991)have examined the hypothalamus and pituitary of individuals affected with autosomal dominant central DI. Braverman et al. (1965) reported an autopsy study of a 49-year-old man who had two sisters with confirmed central DI and an extended family history very suggestive of autosomal dominant central DI. This man was consuming 4- 17 1 of fluid per day (with a similar amount of urine output) as an adult and reported having had polydipsia, polyuria and nocturia since childhood. An autopsy performed after death from a myocardial infarction revealed a small posterior pituitary that was “histologically intact”. His brain was not abnormal except for the hypothalamus. The supraoptic nucleus showed ‘‘a chronic degenerative process consisting of a marked loss of nerve cells, a mild gliosis, and varying degrees of karyorrhexis, chromatolysis, and loss of Nissl substance in the few remaining nerve cells” and the paraventricular nucleus showed similar, but less striking, changes. Green et al. (1967) reported an autopsy study of a 37-year-old woman with a very extensive family history of DI who had had central DI since the age of three. After her death from widely metastatic breast carcinoma, she was found to have a posterior pituitary that was “normal in appearance, but no neurosecretory material was identifiable”. The hypothalamus showed a paucity (“less than 5 % ” ) of the large neuron bodies that normally make up the supraoptic nucleus and an even more complete loss of the large neurosecretory cell bodies, but not the smaller neurons, in the paraventricular nucleus. There was a mild to moderate gliosis of these hypothalamic nuclei. Neurosecretory material was demonstrated in the few remaining large neurons by an aldehyde-fuchsin staining technique. Recently, Bergeron et al. (1991) reported an immunohistochemical study of the hypothalamus of a 72-yearold man with biochemically documented central DI

297

who had three clinically affected sons and one affected grandson. After death from heart failure, examination of the supraoptic nucleus showed virtual absence of magnocellular neurons, but those remaining were non-immunostaining for AVP. A fourth autopsy study (Nagai et al., 1984) did hot demonstrate any histological abnormality in the hypothalamus of a 44-year-old man with polyuria and polydipsia and a family history of DI who died of a myocardial infarction. However, this patient was able to concentrate his urine, produce some AVP, and maintain a nearly normal serum osmolality after a 14-h fluid restriction. Furthermore, the family history of DI included two brothers, a maternal aunt, and the maternal grandfather, but neither his mother nor any of his children or children of his affected siblings were affected. Thus this family had atypical familial polydipsia and polyuria that may represent an extremely mild variant of autosomal dominant central DI, or possibly a distinct disorder. These few histological studies in humans suggested that autosomal dominant central DI results from a dysgenesis or degeneration of the neurons that produce AVP in the supraoptic and paraventricular nuclei of the hypothalamus. The variation in severity of illness over time and between various members of a family may be related to the degree of anatomic and functional abnormality in the hypothalamus. While these findings did not shed light on the genetic basis for the disorder, they suggested a developmental abnormality or possibly a hereditary autoimmune disorder or an abnormal apoptosis (programmed cell death). Initial attempts to define the genetic locus of autosomal dominant central DI were unsuccessful. Fraser (1955) and Pedersen et al. (1985) examined large affected families for linkage (co-inheritance) of this form of DI with various biochemical, phenotypical, or cytogenetic markers such as red blood cell antigens, hair and eye color, HLA type, polymorphic red blood cell enzymes and chromosomal centromeric heteromorphisms. Neither investigator was able to identify significant linkage between DI and any of these markers that are distributed across the human genome.

Molecular biology of the AVP gene

The rat (Schmale et al., 1983) and human (Sausville et al., 1985) genes that encode AVP have been sequenced and have a similar organization (Fig. 1). Riddell et al. (1985) demonstrated that the human gene is located on chromosome 20. Both the rat and human genes are small (2.5 kilobases (kb)) composite genes comprised of three exons that encode three peptides downstream from a signal peptide: AVP, neurophysin I1 (NP 11), and a small glycoprotein (see Van Leeuwen, this volume). The first exon encodes the signal peptide, AVP, and the first nine amino acids of NP 11. The second exon encodes the mid-portion of NP I1 and the third exon encodes the last sixteen amino acids of NP I1 and the small glycoprotein of unknown function. A single mRNA is transcribed from the gene and translated to produce a single precursor protein. Post-translational cleavage yields the component peptides (Gainer et al., 1977). Molecular biology of hereditary central DI in the rat

The mutation that causes hereditary central DI in an animal model, the Brattleboro rat, falls within the AVP-NP I1 gene. This model system has recently been reviewed (Ivell et al., 1990). In the Brattleboro rat, DI is autosomal recessive, in sharp contrast to the human disease. Homozygous Brattleboro rats exhibit poly-dipsia and polyuria due to absence of detectable circulating AVP (Valtin et al., 1974; Pickering and North, 1982) and these rats are also unable to produce NP 11. They can concentrate their urine in response to exogenous AVP, as expected, and they apparently have normal oxytocin production and release from the posterior pituitary (Land

AVP-NP I1

1

2 3

OT-NP I

c-

w

32 1

Fig. 1. Organization of AVP/OT region of human chromosome 20. Arrows indicate orientation of each gene from 5 ' to 3 ' . Numerals and wide line segments indicate exons.

298

et al., 1983; Ivell and Richter, 1984). Sequencing of the Brattleboro rat AVP-NP I1 gene (Schmale and Richter, 1984) revealed a single base pair deletion in the distal portion of the second exon, resulting in a frameshift mutation within the region of the AVPNP I1 gene encoding NP 11. The mutation does not affect transcription of the messenger RNA (mRNA) for AVP-NP 11; however, translation of this mutant mRNA, if it did occur, would produce avery abnormal NP 11. The mutant mRNA encodes a NP I1 with multiple amino acid substitutions in the C-terminus including alteration of an arginine that is probably required to direct proteolytic cleavage of the glycoprotein from the NP I1 and destruction of a glycosylation site. Furthermore, the termination codon is deleted, potentially allowing protein translation to continue abnormally into the mRNA's poly-A tail and thus adding a poly-lysine tail to the protein (Schmale and Richter, 1984; hell et al., 1986). In spite of the completely normal AVPcoding region in this mutant AVP-NP I1 gene, the downstream disruption within NP I1 prevents the successful production of adequate amounts of AVP (Majzoub et al., 1984). The mutant gene does remain capable of being transcribed and translated and the precursor protein appears in the endoplasmic reticulum (ER) but not in the Golgi apparatus or neurosecretory granules (Guldenaar et aI., 1986; Ivell et al., 1986; Krisch et al., 1986). Elegant in vitro experiments by Schmaleet al. (1989) suggest that the N-terminus of the abnormal precursor protein enters the ER but the C-terminus is unable to follow, leaving the precursor anchored in the ER membrane. Neither introducing a stop codon, a glycosylation site, nor a shortened or deleted poly-lysine tail into the mutant mRNA by genetic' engineering techniques allowed normal internalization and processing of the precursor protein in the ER. This suggested that another feature of the mutant NP I1 sequence was responsible for failure of protein processing. Surprisingly, the Brattleboro rats have been found to have a normal amount of AVP of unknown function in their adrenals and ovaries (Lim et al., 1984; Nussey et al., 1984). The only known difference between the

1

1

I

'

I

-

I

I I

Fig. 2. Pedigrees of autosomal dominant central DI families. Solid circles and squares indicate affected females and males, respectively. Open circles and squares indicate unaffected individuals. Slashes indicate deceased individuals. * Signifies affected persons studied in Fig. 3.

AVP-NP I1 mRNA in these tissues and in the hypothalamus is the addition of a relatively short poly-A tail in both the adrenals and ovaries (Ivell et al., 1986). Thus these tissues would produce a mutant precursor with a shorter poly-lysine tail, but the experiments of Schmale et al. (1989) suggest that this alone does not account for the ability of these tissues to produce AVP. In contrast to humans with familial DI, Brattleboro rats have hypertrophied supraoptic and paraventricular nuclei (Sokol and Valtin, 1965).

Molecular biology of human AVP gene in central DI We investigated the possibility that human autosoma1 dominant central DI might be encoded by a mutation in the AVP-NP I1 gene. Our studies involved thirteen affected and eight unaffected members of three multigenerational families (Fig. 2). Affected individuals had otherwise unexplained polyuria and polydipsia since childhood with a urine concentrating response to exogenous AVP or AVP analog and a family history of DI consistent with autosomal dominant inheritance. First, the AVP-NP I1 genes of one affected member of each family were examined for a large

299

Bgl I 1

Pvu I I

kb

-

7.5 6.7

=

5.3 5.0

-

3.8

-

2.3

- 0.8 Fig. 3. Autoradiogram of Southern blot after digestion of genomic DNA with Bgl I1 (left) or Pvu I1 (right) and hybridization with the AVP-NP I1 probe. Lanes I and 2 in each group are unaffected individuals. Lanes3 ,l a n d 5 are affected individuals from families 2, 1, and 3, respectively, as indicated by an asterisk in Fig. 2. Lane 3 in each group contained less DNA than the others, but showed identical patterns upon longer exposure. (From Repaske et al., 1990, with permission from the publisher.)

molecular disruption such as a deletion, insertion or rearrangement (Repaske et al., 1990). Genomic DNA was isolated from peripheral leukocytes (Kunkel et al., 1977) and analyzed by restriction endonuclease digestion followed by Southern blotting (Southern, 1975) and hybridization with a genomic AVP-NP I1 probe. In general, this approach involves digestion of genomic DNA with a restriction enzyme that produces multiple specific DNA fragments of varying sizes. The ends of these fragments are determined by the location of the specific DNA sequence at which that enzyme cuts double-stranded DNA. The resultant collection of DNA fragments is

separated on the basis of size by agarose gel electrophoresis, and the DNA is transferred to a nylon membrane to immobilize the fragments. Fragments that contain some or all of a particular DNA sequence can then be detected by hybridization with a radioactively labeled probe composed of that DNA sequence and by visualization using autoradiography (see Summar, this volume). In these experiments, 5 pg of genomic DNA from three unrelated affected individuals and two controls were digested to completion with the restriction endonuclease Bgl 11, and additional 5 pg aliquots were digested with Pvu 11. After electrophoresis and Southern blotting, the fragments containing portions of the AVP-NP I1 gene were detected by hybridization with a probe containing this entire gene. As shown in Fig. 3, the sizes of the genomic fragments of the AVP-NP I1 gene from affected and control individuals were identical. A large deletion involving the AVP-NP I1 gene would have produced a smaller AVP-NP 11-hybridizing fragment, or, if the deletion included a restriction enzyme recognition site, the total number of visualized fragments would have been decreased. Likewise, an insertion that disrupts the gene would be likely to have increased the size of at least one visualized fragment, and a rearrangement would be likely to have altered the pattern of visualized fragments. The results shown in Fig. 3 indicate that a major structural alteration of the AVP-NP I1 gene was not responsible for the disease in these families. Of course, a small gene disruption, such as the single base deletion of the Brattleboro rat or a base substitution, can have major genetic consequences without being detectable by this methodology (see Van Leeuwen, this volume). Linkage strategy We subsequently studied the AVP-NP I1 locus in autosomal dominant central DI using a genetic linkage strategy. This approach does not depend upon directly identifying or detecting the mutation responsible for the disease. Instead, it is an analysis of co-inheritance of individual AVP-NP I1 alleles

300

and the disease phenotype in affected family members. The goal is to demonstrate that one particular AVP-NP I1 allele, that is, the putative mutant gene, is inherited by all affected family members and not by any unaffected family members. If such an allele can be identified, it provides strong evidence that that allele, or the associated allele of another tightly linked gene, is responsible for the disease. The crucial prerequisite for this type of study is to develop a marker to identify and trace individual alleles through the pedigree. Earlier studies (Fraser, 1955; Pedersen et al., 1985) used panels of convenient cytogenetic or biochemical markers scattered across the genome for their linkage studies of autosoma1 dominant central DI. These investigators hoped that the disease locus would fall close to the locus of one of their markers and that they would therefore detect co-inheritance of that marker and the disease phenotype. However, they were not fortunate enough to have one of their markers close enough to the autosomal dominant central DI locus to detect significant linkage. On the other hand, we were able to develop linkage markers specific for a particular candidate locus, the AVP-NP I1 locus, and thus were able to demonstrate linkage in our families between the AVP region of chromosome 20 and autosomal dominant central DI. We used restriction fragment length polymorphisms (RFLPs) as our markers for the AVP-NP I1 alleles. Scattered throughout the genome, frequently in introns or between genes, are single base pair variations that have no effect on phenotype. Such variations - polymorphisms - are usually undetected, but can easily be detected if they fall within a restriction endonuclease recognition sequence as they then produce a polymorphic restriction site (Botsteinet al., 1980; Antonarakis et al., 1982). The presence of a polymorphic restriction site creates an RFLP, a variation in the size of fragments visualized after a restriction digest. Genomic DNA digested with the appropriate restriction enzyme will yield one large fragment (usually designated the - RFLP allele) if the polymorphic restriction siteis absent, or two smaller DNA fragments if the restriction site is present. Usually only one of the smaller fragments

hybridizes with the radiolabeled probe and is visualized. This smaller fragment is usually designated the + RFLP allele. Occasionally, two different polymorphic sites can occur within one large restriction fragment, producing three different restriction fragment lengths. These RFLP alleles are then named by their fragment sizes. The presence of different polymorphic restriction sites, detected as different RFLP alleles, on each of a person’s two homologous chromosomes can serve as a marker to distinguish the .two chromosomes. Identifying a set of RFLPs in a given region of a chromosome allows construction of an RFLP haplotype associated with that region of each chromosome. An RFLP haplotype frequently allows informative analysis of an element of a pedigree where the information from a single RFLP allele might be ambiguous. We first attempted to find RFLPs associated with the AVP-NP I1 gene that could be detected with the AVP-NP I1 DNA probe. Genomic DNA from nine control individuals (representing 18 AVP-NP I1 alleles) was digested with each of 33 restriction endonucleases and analyzed by Southern blotting using the AVP-NP 11 probe. All of the resulting fragments from each enzyme were identical in size in all of the individuals examined (Repaske et al., 1990). In other words, no RFLPs were detected using these restriction enzymes and the AVP-NP I1 probe. We then used a DNA probe from an adjacent gene to identify RFLPs that are very tightly linked to, but not in, the AVP-NP I1 gene. The gene that encodes human oxytocin (OT) and its carrier protein, neurophysin I (NP I), is located adjacent (approximately 10 kb away) to the AVP-NP I1 gene (Fig. 1). At this distance, the likelihood that these two loci would recombine during meiosis is less than 1 in 10000. Thus the AVP-NP I1 allele and the OT-NP I allele on each chromosome 20 are inherited together 9999 times out of 10000.The OT-NP I gene is highly homologous to the AVP-NP I1 gene but is oriented in the opposite direction. It has been cloned by Sausvilleet al. (1985). We used an OT-NP I genomic clone, radioactively labeled, to rescreen the restriction enzyme digested DNA from the nine control in-

301 Xba I

Apa I

21/9 21/3 9/3

3/3

-/-

- / + +/+

Dde I

kb

--/+

+I+

21 9

5

3

kb 1.1

0.7 0.6

0.9

Fig. 4. RFLPs in the AVP/OT region of human chromosome 20. Autoradiograms of Southern blots of control DNA after digestion with indicated restriction endonuclease and hybridization with the OT-NP I probe. RFLP alleles present in each sample are indicated above each lane. Apa I alleles are labeled by approximate fragment size in kb. Xba I and Dde I alleles are labeled (for absence of the polymorphic restriction site) for the larger fragment and + (for presence of the polymorphic restriction site) for the smaller fragment. The Dde I digest reveals a non-polymorphic fragment of 1.1 kb in association with the RFLP. (From Repaske et al., 1990, with permission from the publisher.)

TABLE 1 OT-NP I RFLP allele sizes and frequencies Enzyme

Allele desig- Size (kb) Frequency Chromosomes nation examined

Apa I

21 9 3

XbaI

-

DdeI

-

sea I

-

+

+ +

21.4 9.4 3.1

0.21 0.40 0.39

62

9.8 5.4

0.92 0.08

100

0.66 0.60

0.07 0.93

58

0.90 0.10

18

17.5 15.5

dividuals (Repaske et al., 1990). Four different restriction enzymes (Apa I, Xba I , Dde I, and Sca I) revealed RFLPs with the OT-NP I probe (Fig. 4, Table I ) . For example, there is a polymorphic XbaI site adjacent to the OT-NP I gene that falls between two constant Xba I sites, 9.8 kb apart, that flank this gene. Thus, the OT-NP I probe detects a 9.8 kb fragment if the polymorphic Xba I site is absent but detects a 5.4 kb fragment in the presence of the site. Evaluation of additional control individuals for these RFLPs allowed determination of the frequency of occurrence of each fragment length in the control population (Table I). Five microgram aliquots of genomic DNA from each available member of the three affected families were then digested separately with Xba I (Fig. 5), Apa I (Fig. 6 ) , and Dde I. In a few cases, not all digestions could be performed because of limited DNA quantity. The pattern of OT-NP I RFLPs in each individual allowed construction of RFLP haplotypes that demonstrate the inheritance pattern of individual OT-NP I (and therefore the tightly linked AVP-NP 11) alleles for families 1 and 2 (Fig. 7). In family 1, the haplotype (Apa I/Dde I/Xba I ) 3/ + / + occurs in all affected persons and not in unaffected persons. Likewise, in family 2, the disease is associated with the 3/ - / - haplotype. Unfortunately, all members of family 3 were homozygous for all three of these RFLPs as well as the Sca I RFLP. Thus these RFLPs were unable to distinguish the various OT-NP I alleles in this family. The statistical significance of co-inheritance of the RFLP haplotype and the DI phenotype, taking into account such variables as the RFLP allele frequencies, was calculated as a LOD (log odds) score (Morton, 1955). This analysis allows data from different families with the same disease to be combined in order to achieve greater statistical significance than with any one family alone. The maximum LOD score for family 1 alone was 2.40 which means that there is only a 1 in 102.4,or 1 in 250, probability that this association between the disease and the RFLP haplotype in the AVP/OT region of chromosome 20 occurred by chance alone. Combining the data from the families 1 and 2 raises the LOD score to 2.7, or

302

kb

9

w

9.89

5.4Fig. 5 . Autoradiograms of genomic DNA from members of families 1 (left) and 2 (right) after digestion with Xba I, Southern blotting, and hybridization with the OT-NP I probe. (From Repaske et al., 1990, with permission from the publisher.)

-2 1 - 9

Fig. 6. Autoradiograms of genomic DNA from members of families 1 (left) and 2 (right) after digestion with Apa I, Southern blotting, and hybridization with the OT-NP I probe, (From Repaske et al., 1990, with permission from the publisher.)

c

303

Apa I

Dde I Xba I

Fig. 7. RFLP allele haplotypes of individuals from families 1 (top) and 2 (bottom Symbols representing the haplotypes of RFLPs on each chromosome are displayed vertically. The RFLP alleles were detected by hybridization to the OT-NP I probe after digestion with (from top to bottom) Apa I, Dde I, and Xba I. DNA was not available from the deceased members of family 1 or from the members of family 2 lacking haplotype symbols. Insufficient quantities of DNA prevented determination of complete haplotypes in all cases. (From Repaske et al. 1990, with permission from the publisher.)

a 1 in 500 probability that these results occurred by chance alone. Recently we have studied DNA from a fourth unrelated family that contains 11 affected individuals. Preliminary results in five informative matings substantiate the findings that the AVP/OT locus cosegregates with the DI phenotype. Sequence analysis of AVP gene

These results very strongly suggest that the genetic locus for autosomal dominant central DI is in or near the AVP-NP I1 gene as in the Brattleboro rat. A recent study by Ito et al. (1991) examined the nucleic acid sequence of the AVP-NP I1 genes of two affected sisters from a Japanese family with this disease. They identified a single base pair alteration in one allele of both of these patients compared with three Japanese and two Caucasian controls and with the previously published gene sequence (Sausville et al., 1985). This sequence difference is a G to A transition located in the NP I1 coding sequence in exon 2, just 21 base pairs 5 ’ to the corresponding site of the Brattleboro rat deletion. As a result of this alteration a glycine (position 57) in the NP I1 is

changed to a serine. Unfortunately, this study could not include any unaffected family members to demonstrate that this alteration is present only in affected individuals. Thus it remains possible that this alteration is a phenotypically silent polymorphism in this family and that the disease is encoded by a mutation at a different locus. The G to A transition in this Japanese family destroys an Msp I restriction endonuclease recognition site (CCGG to CCAG). This allowed relatively rapid screening of individuals for the alteration. Ito et al. designed specific polymerase chain reaction (PCR) primers that flank exon 2 of the AVP-NP I1 gene and allow amplification of a 304 base pair (bp) fragment including all of exon 2. There are three Msp I sites within this fragment in control individuals, and therefore digestion of this PCR product with Msp I produces four fragments of 32 bp, 56 bp, 93 bp and 123 bp. One of the two AVP-NP I1 alleles of each of the affected individuals in the Japanese family was missing the Msp I site between the 123 and 93 bp fragments, and thus these individuals have a 216 bp fragment in addition to the fragments from their normal allele. No 216 bp frag-

304

ments were detected in a screening of 20 unrelated, unaffected Japanese individuals, providing additional evidence that this alteration is not simply a polymorphism and may be responsible for autosoma1 dominant central DI in this family. We have recently examined two unaffected and four affected members of our family 1 by a similar PCR amplification of AVP-NP I1 exon 2 and did not find that this Msp I site is destroyed in any of these individuals, suggesting heterogeneity in the molecular basis of autosomal dominant central DI. Further studies of the AVP-NP I1 gene in this and other affected families are in progress. An additional unique single base substitution (G T at nucleotide position 1740causing a Glu --* Val substitution at position 29) in the coding region for neurophysin I1 has recently been shown to be associated with a family affected by autosomal dominant central DI (Schmale et al., 1991). Further studies have suggested that this mutation is the cause of autosomal dominant central DI in this family (see Van Leeuwen, this volume). +

Summary and conclusions Molecular biology techniques have begun to shed light on the genetic basis of autosomal dominant central DI, but several very basic questions remain to be answered. The disorder was initially presumed to have a developmental, degenerative, or autoimmune basis based on the autopsy findings in the hypothalamus of a limited number of patients. The molecular cloning of the AVP-NP I1 gene and the clue from the Brattleboro rat that at least this one form of hereditary DI involved an AVP-NP I1 gene mutation allowed us to focus on this gene in our study of human hereditary DI. Our initial experiments did not show this gene to have a major structural alteration such as a deletion, insertion, or rearrangement, but the approach was not capable of detecting more subtle defects. The linkage studies provided substantial evidence that one particular OT-NP I haplotype was linked to the disease phenotype in each family, and thus, a mutation in the AVP/OT region of chromosome 20 is responsible

for this disease. Ito et al. (1991) then identified a single base change in the AVP-NP I1 gene in affected members of one Japanese family. This change was not detected in unrelated, unaffected persons and thus is a good candidate for the mutation causing the disease in this family. However, there appears to be diversity in the molecular basis of autosomal dominant central DI as affected members of one of our families did not have this particular base change in either AVP-NP I1 allele and recently another distinct AVP-NP I1 gene base change has been associated with this disorder. One interesting question still to be addressed is how a mutation in the NP-I1 coding region of this gene prevents AVP release from the posterior pituitary in the rat or the human disease. Does the disrupted AVP-NP I1 coding sequence prevent normal processing of the mRNA so that it can not be properly translated into protein? Does the mutated AVP-NP I1 glycoprotein precursor protein interfere with normal post-translational processing to prevent release of AVP? Is an altered NP I1 protein not able to protect the AVP from proteolysis within the magnocellular neuron? An even more puzzling question is how a mutation in the gene encoding a hormone is inherited in an autosomal dominant pattern. The Brattleboro rat model follows the a priori expectation of autosomal recessive inheritance: the animal only exhibits a defect in hormone function if both genes encoding the hormone are defective. In the human disease, the dysfunction of one gene somehow inhibits the normal functioning of the other AVP-NP I1 gene. The mutant gene product may irreversibly bind the ribosome blocking further AVP synthesis, even that encoded by the normal gene. The degenerative changes of the cells that produce AVP in autosomal dominant central DI could be the result of a gene product that is somehow toxic, although this would not easily explain the variable expression of the disease over time and the variability from one affected family member to another. Thus while the molecular genetic approach has allowed us to begin to elucidate the molecular basis of autosomal dominant central DI, many unanswer-

305

ed questions still exist about the pathophysiology of this disorder. Sequencing of the mutant gene in additional families and then study of mRNA synthesis, protein synthesis, and subsequent protein processing from these mutant genes should allow a definitive answer to the questions raised above and contribute to a fuller understanding of the molecular basis of autosomal dominant central DI.

References Antonarakis, S.E., Phillips, J.A., 111and Kazazian, H.H. (1982) Genetic diseases: diagnosis by restriction endonuclease analysis. J. Pediatr., 100: 845 - 856. Bergeron, C., Kovacs, K., Ezrin, C. and Mizzen, C. (1991) Hereditary diabetes insipidus: an immunohistochemical study of the hypothalamus and pituitary gland. Acta Neuropathol. (Bed.), 81: 345 - 348. Blackett, P.R., Seif, S.M., Altmiller, D.H. and Robinson, A.G. (1983) Familial central diabetes insipidus: vasopressin and nicotine stimulated neurophysin deficiency with subnormal oxytocin and estrogen stimulated neurophysin. Am. J. Med. Sci., 286: 42-46. Bode, H.H. andcrawford, J.D. (1969) Nephrogenic diabetes insipidus in North America - The Hopewell hypothesis. N . Engl. J. Med., 280: 750- 754. Botstein, D., White, R.L., Skolnick, M. andDavis, R.W. (1980) Construction of a genetic linkage map in man using restriction fragment length polymorphisms. A m . J. Hum. Genet., 32: 314-331. Braverman, L.E., Mancini, J.P. and McGoldrick, D.M. (1965) Hereditary idiopathic diabetes insipidus. A case report with autopsy findings. Ann. Int. Med., 63: 503 -508. Brown, D. and Orci, L. (1983) Vasopressin stimulates formation of coated pits in rat kidney collecting ducts. Nature, 302: 253 - 255. Brownstein, M. J. (1983) Biosynthesis of vasopressin and oxytocin. Annu. Rev. Physiol., 45: 129- 135. Dousa, T.P., Barnes, L.D. and Kim, J.K. (1977) The role of cyclic AMP-dependent protein phosphorylations and microtubules in the cellular action of vasopressin in mammalian kidney. In: A.M. Moses and L. Share (Eds.), Neurohypophysis, Karger, Basel, pp. 220- 235. Forssman, H. (1955) Two different mutations of the Xchromosomecausing diabetes insipidus. Am. J. Hum. Genet., 7: 21 -27. Fraser, F.C. (1955) Hereditk dominante d’un diabete insipide du a une dkficience en pitressine. J. Gener. Hum., 4: 193 - 203. Gainer, H., Same, Y. and Brownstein, M.J. (1977) Biosynthesis and axonal transport of rat neurohypophyseal proteins and peptides. J. Cell Biol., 13: 366-381.

Green, J.R., Buchan, G.C., Alvord, E.C. and Swanson, A.G. (1967) Hereditary and idiopathic types of diabetes insipidus. Brain, 90: 707-714. Guldenaar, S.E.F., Nahke, P. and Pickering, B.T. (1986) Immunocytochemical evidence for the presence of a mutant vasopressin precursor in the supraoptic nucleus of the homozygous Brattleboro rat. Cell TissueRes., 244: 431 - 436. Herbert, S.C. and Andreoli, T.E. (1984) Effects of antidiuretic hormone on cellular conductive pathways in mouse medullary thick ascending limb of Henle. 11. Determinants of the ADHmediated increases in transepithelial voltage in net C1- absorption. J. Membr. Biol., 80: 221 -233. Iovino, M. and Steardo, L. (1984) Vasopressin release to central and peripheral angiotensin I1 in rats with lesions of the subfornical organ. Brain Res., 322: 365 - 368. Ito, M., Mori, Y., Oiso, Y. and Saito, H. (1991) A single base substitution in the coding region for neurophysin I1 associated with familial central diabetes insipidus. J. Clin. Invest., 87: 725 - 728. Ivell, R. and Richter, D. (1984) Structure and comparison of the oxytocin and vasopressin genes from rat. Proc. Natl. Acad. Sci. U.S.A., 81: 2006-2010. Ivell, R., Schmale, H., Krisch, B., Nahke, P. and Richter, D. (1986) Expression of a mutant vasopressin gene: differential polyadenylation and read-through of the mRNA 3’ end in a frameshift mutant. EMBO J., 5: 971 - 977. Ivell, R., Burbach, J.P.H. and Van Leeuwen, F.W. (1990) The molecular biology of the Brattleboro rat. Front. Neuroendocrinol., 11: 313 - 338. Jard, S., Roy, C., Barth, T., Rajerison, R. and Bockaert, J. (1975) Antidiuretic hormone-sensitive kidney adenylate cyclase. Adv. Cyclic Nucleotide Res., 5: 31 - 52. Kaplowitz, P.B., D’Ercole, A.J. and Robertson, G.L. (1982) Radioimmunoassay of vasopressin in familial central diabetes insipidus. J. Pediatr., 100: 76 - 81. Krisch, B., Nahke, P. and Richter, D. (1986) Immunocytochemical staining of supraoptic neurons from homozygous Brattleboro rats by use of antibodies against two domains of the mutated vasopressin precursor. Cell Tissue Res., 244: 351 - 358. Kunkel, L.M., Smith, K.D., Boyer, S.H., Borgaonkar, D.S. and Wachtel, S.S. (1977) Analysis of human Y chromosome specific reiterated DNA in chromosome variants. Proc. Natl. Acad. Sci. lJ,S.A., 14: 1245- 1249. Land, H., Grez, M., Ruppert, S., Schmale, H., Rehbein, M., Richter, D. and Schutz, G. (1983) Deduced amino acid sequence from the bovine oxytocin-neurophysin I precursor cDNA. Nature, 302: 342 - 344. Levinger, E.L. and Escamilfa, R.F. (1955) Hereditary diabetes insipidus: report of 20 cases in seven generations. J. Clin. Endocrinol. Metab., 15: 547- 552. Lim, A.T.W., Lolait, S. J., Barlow, J.W., Autelitano, D.J., Toh, B.H., Boublik, J., Abraham, J., Johnston, C.I. and

306 Funder, J.W. (1984) Immunoreactive arginine-vasopressin in Brattleboro rat ovary. Nature, 310: 61 -64. Majzoub, J.A., Pappey, A., Burg, R. and Habener, J.F. (1984) Vasopressin gene is expressed at low levels in the hypothalamus of theBrattleboro rat. Proc. Natl. Acad. Sci. U.S.A.,81: 5296- 5299. Martin, F.I.R. (1959) Familial diabetes insipidus. J. Med., 112: 573 - 582. Meinders, A.E. and Bijlsma, J.B. (1970) A family with congenital hypothalamic neurohypophyseal diabetes insipidus. Folia Med. Neerl., 13: 68 - 72. Moehlig, R.C. and Schultz, R.C. (1955) Familial diabetes insipidus. Report of one of fourteen cases in four generations. JAMA, 158: 725-727. Morton, N.E. (1955) Sequential tests for the detection of linkage. A m . J. Hum. Genet., 7: 277-318. Nagai, I., Li, C.H., Hsieh, S.M., Kizaki, T. and Urano, Y. (1984) Two cases of hereditary diabetes insipidus, with an autopsy finding in one. Acta Endocrinol., 105: 318 - 323. Nussey, S.S., Ang, V.T.Y., Jenkins, J.S., Chowdrey, K.S. and Bisset, G.W. (1984) Brattleboro rat adrenal contains vasopressin. Nature, 310: 64 -66. Pedersen, E.B., Lamm, L.U., Albertsen, K., Madsen, G., Bruun-Petersen, G., Henningsen, K., Friedrich, U. and Magnusson, K. (1985) Familial cranial diabetes insipidus: a report of five families. Genetic, diagnostic and therapeutic aspects. Q. J. Med., 224: 883 - 896. Pender, C.B. and Fraser, F.C. (1953) Dominant inheritance of diabetes insipidus. A family study. Pediatrics, 11: 246- 254. Pickering, B.T. and North, W.G. (1982) Biochemical and functional aspects of magnocellular neurons and hypothalamic diabetes insipidus. Ann. N . Y . Acad. Sci., 394: 72-81. Repaske, D.R., Phillips, J.A., 111, Kirby, L.T.,Tze, W.J., D’ErCole, A.J. and Battey, J . (1990) Molecular analysis of autosomal dominant neurohypophyseal diabetes insipidus. J. Clin. Endocrinol. Metab., 70: 752 - 757. Reeves, W.B. and Andreoli, T.E. (1989) Nephrogenic diabetes insipidus. In: C.R. Scriver, A.L. Beaudet, W.S. Sly and D. Valle (Eds.), Metabolic Basis of Inherited Disease, 6th edn., McGraw-Hill, New York, pp. 1985 - 201 1. Riddell, D.C., Mallonee, R., Phillips, J.A., 111, Parks, J.S., Sexton, L.A. and Hamerton, J.L. (1985) Chromosomal assignment of human sequences encoding arginine vasopressinneurophysin I1 and growth hormone releasing factor. Somat. Cell, Mol. Genet., 11: 189- 195. Sausville, E., Carney, D. and Battey, J . (1985) The human vasopressin gene is linked to the oxytocin gene and is selectively expressed in a cultured lung cancer cell line. J. Biol. Chem., 260: 10236- 10241. Sawchenko, P.E. and Swanson, L.W. (1981) Central noradrenergic pathways for the integration of hypothalamic neuroendocrine and autonomic responses. Science, 214: 685 - 687. Schmale, H. and Richter, D. (1984) Single base deletion in the

vasopressin gene is the cause of diabetes insipidus in Brattleboro rats. Nature, 308: 705 -709. Schmale, H., Heinsohn, S. and Richter, D. (1983) Structural organization of the rat gene for the arginine vasopressinneurophysin precursor. EMBO J., 2: 763 - 767. Schmale, H., Borowiak, B., Holtgreve-Grez, H. and Richter, D. (1989) Impact of altered protein structures on the intracellular traffic of a mutated vasopressin precursor from Brattleboro rats. Eur. J. Biochem., 182: 621 -627. Schmale, H., Bahnsen, U., Fehr, S., Nahke, P. and Richter, D. (1991) Hereditary diabetes insipidus in man and rat. In: S. Jard and R. Jamison (Eds.), Vasopressin - Colloque INSERM, Vol. 208, John Libbey Eurotext Ltd., pp. 57 - 62. Schrier, R.W., Berl, T. and Anderson, R.J. (1979) Osmotic and non-osmotic control of vasopressin release. Am. J. Physiol., 236: F321- F332. Sokol, H.W. and Valtin, H. (1965) Morphology of the neurosecretory system in rats homozygous and heterozygous for hypothalamic diabetes insipidus (Brattleboro strain). Endocrinology, 77: 692 - 700. Southern, E.M. (1975) Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol., 98: 503 - 517. Toth, E.L., Bowen, P.A. and Crockford, P.M. (1984) Hereditary central diabetes insipidus: plasma levels of antidiuretic hormone in a family with a possible osmoregulator defect. Can. Med. Assoc. J . , 131: 1237- 1241. Valtin, H., Stewart, J. and Sokol, H.W. (1974) Genetic control of the production of posterior pituitary principles. Handbk. Physiol., 7: 131- 171. Zerbe, R.L. and Robertson, G.L. (1987) Osmotic and nonosmotic regulation of thirst and vasopressin secretion. In: M.H. Maxwell, C.R. Kleeman and R.G. Narins (Eds.), Clinical Disorders of Fluid and Electrolyte Metabolism, 4th edn., McGraw-Hill, New York, pp. 61 - 78.

Discussion D.F. Swaab: How can we explain that in human hereditary diabetes insipidus (DI) the vasopressin cells in the supraoptic and paraventricular nuclei (but probably not in the suprachiasmatic nucleus) die off (Bergeron et al., 1991)but not in the Brattleboro rat? D.R. Repaske: This is a good question and the answer is not yet clear. The Brattleboro rat has hypertrophied supraoptic and paraventricular nuclei and posterior pituitary (Sokol and Valtin, 1965). All of these structures are relatively atrophic in the human disease (Braverman et al., 1965; Green et al., 1967; Bergeron et al., 1991). This difference must reflect different pathophysiological mechanisms which cause the disease in the rat and the human. The mechanisms must be very different as the heterozygous state in the rat produces a nearly normal phenotype while

307 in the human the disease is manifest (see also, Van Leeuwen, this volume). Perhaps the mutant precursor accumulates in neurons in the human disease in a manner that is toxic to these cells or in some manner that leaves them susceptible to autoimmune attack. The selective atrophy of the AVP-containing neurons in the supraoptic and paraventricular nuclei is reminiscent of the changes seen after high stalk section (MacCubbin and Van Buren, 1963; Morton, 1969) and may reflect a primary lesion (such as autoimmune attack or toxic degeneration) which takes place in the neurohypophyseal nerve terminals. F.W. van Leeuwen: With respect to the point of the explanation of cell loss in the supraoptic and paraventricular nucleus (SON and PVN) in human hypothalamic DI, it might be that superactivation of vasopressin (VP) cells leads to cell death whereas hyperactivity results in a prolonged life as compared to hypoactivity. D.F. Swaab: I agree that even if the “use it or lose it” hypothesis will be proven to be correct, it will only hold within certain physiological limits. Extreme stimulation by excitoxin administration for instance has shown to lead to cell death (Swaab, 1991a,b). For the hypothalamic diabetes insipidus (DI) Brattleboro rat it has been repeatedly shown that the SON and PVN are hyperactive(Soko1 and Valtin, 1965; Swaab et al., 1973; Van Tol, 1987; Sherman et al., 1988). Although nobody, t o my knowledge, has counted the SON and PVN cells in the homozygous DI rat, we have no reason to assume that in these nuclei cell loss in any considerable degree occurred. Nor is there any reason to assume that in the human DI the SON and PVN cells should be much more stimulated than in the DI rat (what you called superstimulation). Therefore I prefer the idea that some compound produced or accumulated by the disease process in the human DI neurons is causing their degeneration rather than their hyperactivity. J.B. Martin: Do you have any explanation for how a point mutation in one allele can cause an autosomal dominant disorder? Why does not the normal allele generate sufficient AVP t o give rise to a partial deficiency? Is it possible that there is adequate vasopressin secretion early in life and that the secretory capacity fails with advancing age? D.R. Repaske: We do not yet know the pathophysiological mechanism of the human disease. Expression of the human mutant allele, in contrast to the Brattleboro rat mutant allele, must create a significant block to expression of the normal allele. The mutant allele may cause the death of AVP-producing neurons by toxic or autoimmune mechanisms or it may otherwise interfere with a critical step in the production of the normal hormone. Possibilities include an inability of the mutant precursor protein to release from the ribosome, from a transport system in the ER/Golgi apparatus, or from a critical post-translational processing enzyme. Reply to J.B. Martin by F.W. Van Leeuwen: Most probably a very low amount of AVP is present in the plasma of human hypothalamic DI persons since chlorpropamide only acts in the presence of AVP (Miller and Moses, 1970). This makes it likely

that some AVP is released from the hypothalamo-neurohypophyseal system. This indicates that the normal AVP allele is expressed and that some is translated, packaged, transported and released. This is different from the DI + / + rat in which no AVP is detectable in the plasma and chlorpropamide has no effect (Berndt et al., 1970). Furthermore it has been shown by Schmale et al. (1989) that in the homozygous Brattleboro rat the cause of the arrested transport from the endoplasmic reticulum to the Golgi apparatus of the altered AVP precursor is most probably a change in the internal domain of the precursor. In the human situation a similar phenomenon may occur where the slightly different AVP precursor may disturb the intracellular transport or packaging of the wild-type precursor (see Van Leeuwen, this volume). D.R. Repaske: Kaplowitz et al. (1982) also reported that AVP was detectable in low amounts in an individual with central DI, while his affected brother had essentially no detectable AVP. Also, case reports indicate that the degree of phenotypic expression of DI varies with age. This is likely due to variations in the amount of AVP production during life. Thus, while the mutation creates sufficient dysfunction in AVP production to cause the clinical disease, the capacity to make some AVP is retained. D.R. Repaske: Failure of DNA repair mechanisms in the AVPNP I1 gene is an unlikely explanation for autosomal dominant central DI as: (1) some affected individuals express the disease from birth; (2) some affected individuals clinically improve in later life; (3) there appears to be a single consistent AVP-NP 11 gene mutation within each family where a mutation has been identified; and (4) failure of a repair mechanism might be expected to affect either AVP-NP I1 allele with equal frequency; however, our linkage analysis demonstrates that the disease is consistently associated with one particular AVP-NP I1 allele in each family.

References Bergeron, C., Kovacs, K., Ezrin, C. and Mizzen, C. (1991) Hereditary diabetes insipidus: an immunohistochemical study of the hypothalamus and pituitary gland. Acta Neuropathol. (Bed.), 81: 345 - 348. Berndt, W.O., Miller, M., Kettyle, W.M. and Valtin, H. (1970) Potentiation of the antidiuretic effect of vasopressin by chlorpropamide. Endocrinology, 86: 1028- 1032. Braverman, L.E., Mancini, J.P. and McGoldrick, D.M. (1965) Hereditary idiopathic diabetes insipidus. A case report with autopsy findings. Ann. Int. Med., 63: 503 - 508. Green, J.R., Buchan, G.C., Alvord, E.C. and Swanson, A.G. (1967) Hereditary and idiopathic types of diabetes insipidus. Brain, 90: 707 - 714. Kaplowitz, P.B., D’Ercole, A.J. and Robertson, G.L. (1982) Radioimmunoassay of vasopressin in familial central diabetes insipidus. J. Pediatr., 100: 76-81. MacCubbin, D.A. and Van Buren, J.M. (1963) A quantitative

308

evaluation of hypothalamic degeneration and its relation to diabetes insipidus following interruption of human hypophyseal stalk. Brain, 86: 443 - 464. Miller, M. and Moses, A.M. (1970) Mechanism of chlorpropamide action in diabetes insipidus. J. Clin. Endocrinol. Mefab., 30: 488 - 496. Morton, A. (1969) A quantitative analysis of the normal neuron population of the hypothalamic magnocellular nuclei in man and of their projections to the neurohypophysis. J. Comp. Neurol., 136: 143- 158. Schmale, H., Bahnsen, U., Fehr, S., Nahke, P. and Richter, D. (1991) Hereditarydiabetes insipidusin manand rat. In: S. Jard and R. Jamison (Eds.), Vasopressin, John Libbey, London, pp. 57 - 62. Sherman, T.G., Day, R., Civielli, O., Douglas, J . , Herbert, E., Akil, H. and Watson, S.J. (1988) Regulation of hypothalamic magnocellular neuropeptides and their mRNAs in the Brattleboro rat: coordinate responses to further osmotic challenge.

J. Neurosci., 8: 3785 - 3796. Sokol, H.W. and Valtin, H. (1965) Morphology of the neurosecretory system in rats homozygous and heterozygous for hypothalamic diabetes insipidus (Brattleboro strain). Endocrinology, 77: 692 - 700. Swaab, D.F. (1991a) Brainagingand Alzheimer’s disease, “wear and tear” versus “use it or lose it”. Neurobiol. Aging, 12: 317-324. Swaab, D.F. (1991b) Author’s response to commentaries. Neurobiol. Aging, 12: 352- 355. Swaab, D.F., Boer, G.J. and Nolten, J.W.L. (1973) The hypothalamo-neurohypophyseal system (HNS) of the Brattleboro rat. Acta Endocrinoi., Suppl., 177: 80. Van Tol, H.H.M. (1987) Regulation of Vasopressin and Oxytocin GeneExpression in the Hypofhalamo-Neurohypophyseal System of the Rat. Dissertation, State University of Utrecht, Utrecht, The Netherlands, p. 129.