Ectopic Mineralization and Conductive Hearing Loss in Enpp1asj Mutant Mice, a New Model for Otitis Media and Tympanosclerosis

Otitis media (OM), inflammation of the middle ear, is a common cause of hearing loss in children and in patients with many different syndromic diseases. Studies of the human population and mouse models have revealed that OM is a multifactorial disease with many environmental and genetic contributing factors. Here, we report on otitis media-related hearing loss in asj (ages with stiffened joints) mutant mice, which bear a point mutation in the Enpp1 gene. Auditory-evoked brainstem response (ABR) measurements revealed that around 90% of the mutant mice (Enpp1asj/asj) tested had moderate to severe hearing impairment in at least one ear. The ABR thresholds were variable and generally elevated with age. We found otitis media with effusion (OME) in all of the hearing-impaired Enpp1asj/asj mice by anatomic and histological examinations. The volume and inflammatory cell content of the effusion varied among the asj mutant mice, but all mutants exhibited a thickened middle ear epithelium with fibrous polyps and more mucin-secreting goblet cells than controls. Other abnormalities observed in the Enpp1 mutant mice include over-ossification at the round window ridge, thickened and over-calcified stapedial artery, fusion of malleus and incus, and white patches on the inside of tympanic membrane, some of which are typical symptoms of tympanosclerosis. An excessive yellow discharge was detected in the outer ear canal of older asj mutant mice, with 100% penetrance by 5 months of age, and contributes to the progressive nature of the hearing loss. This is the first report of hearing loss and ear pathology associated with an Enpp1 mutation in mice. The Enpp1asj mutant mouse provides a new animal model for studying tympanosclerotic otitis and otitis media with effusion, and also provides a specific model for the hearing loss recently reported to be associated with human ENPP1 mutations causing generalized arterial calcification of infancy and hypophosphatemic rickets.


Introduction
While many genes have been discovered that underlie Mendelian forms of sensorineural hearing loss (http://hereditaryhearingloss.org/), most cases of conductive hearing loss, often caused by otitis media and tympanosclerosis, have complex and poorly understood etiologies. Otitis media is an inflammation of the middle ear and is among the most common childhood illnesses. Heritability studies have shown that genetic factors can play an important role in otitis media susceptibility, but few contributing genes have been identified in human populations [1]. In contrast to human studies, a growing number of mouse mutations have been identified that manifest a high incidence of otitis media, including Eya4, Tlr4, p73, MyD88, Fas, E2f4, Plg, Fbxo11, Evi1 [1,2], Sh3pxd2b [3], Rpl38 [4], Isl1 [5], Chd7 [6], Lmna1 [7], Phex [8], Oxgr1 [9], Tgif1 [10], and Mcph1 [11]. The wide diversity of these genes and their mutant pathologies, including craniofacial abnormalities with Eustachian tube malformations and innate immune response defects, underscores the complex nature of otitis media.
Here we report on the conductive hearing loss associated with a recessive ENU-induced missense mutation of the ectonucleotide pyrophosphatase/phosphodiesterase 1 gene (Enpp1) that was discovered at The Jackson Laboratory and named "ages with stiffened joints" (asj) because of the progressive ankylosis and osteoarthritis exhibited by mutant mice [12]. The Enpp1 gene encodes an enzyme (ENPP1) that regulates soft-tissue calcification and bone mineralization by producing inorganic pyrophosphate, a major inhibitor of calcification. A previous analysis of asj mutant mice found that enzymatic activity of ENPP1 in the liver and inorganic pyrophosphate (PPi) levels in plasma were markedly reduced [13]. Other mouse mutations of Enpp1 have been reported, including the naturally occurring "tiptoe walking" (ttw) nonsense mutation [14], an ENU-induced C397S missense mutation [15], and a genetically engineered knockout mutation [16]. The Enpp1 mutant mice in these studies exhibited extensive mineralization defects in a number of tissues, including spinal ligaments, long bones, articular cartilage, heart, aorta, arterial blood vessels, vibrissae, liver, kidneys, and retina. The effects of the Enpp1 mutations on auditory function and middle and inner ear histology of mice, however, were not examined.
We show here that Enpp1 asj mutant mice exhibit a conductive hearing loss that is associated with middle and inner ear mineralization abnormalities. These mice provide a new animal model for studies of otitis media and tympanosclerosis and for the hearing loss recently shown to be associated with human ENPP1 mutations causing generalized arterial calcification of infancy and hypophosphatemic rickets [17,18].

Mice
The recessive asj mutation was discovered at The Jackson Laboratory in 2004 in the C57BL/6J progeny of an ENU treated C57BL/6J male. The mutant strain name is C57BL/6J-Enpp1 asj / GrsrJ and is available from The Jackson Laboratory (Stock #012810). Experimental mice were housed in the Research Animal Facility of The Jackson Laboratory, and all procedures involving their use were approved by the Institutional Animal Care and Use Committee. The Jackson Laboratory is accredited by the American Association for the Accreditation of Laboratory Animal Care.

Genotyping the Enpp1 asj mutation
The asj mutation is a single nucleotide substitution of T to A in exon 7 of the Enpp1 gene resulting in an amino acid substitution from valine to aspartic acid at residue 246 (p.V246D) of protein Reference Sequence NP_032839. The T to A point mutation (nt 848 of mRNA Reference Sequence NM_008813) creates a new TaqI restriction site (TCGA), which is the basis of the genotyping method described by Harris and colleagues [12] and that was used in this study to distinguish wild-type from mutant alleles.

Assessment of hearing by ABR
Hearing was evaluated in anesthetized (Avertin, 0.4 mg/g mouse mass) wild-type, Enpp1 asj/+ , and Enpp1 asj/asj mice by measuring auditory brainstem response (ABR). A computer-aided evoked potential system (Intelligent Hearing Systems) was used to measure mouse ABR thresholds as previously described [19].

Histological analyses of middle and inner ears
Histological analyses of the middle and inner ears were performed following the methods described previously [20]. Briefly, middle and inner ears from Enpp1 asj/asj mice and wild-type mice were dissected after transcardial perfusion with Bouin's fixative. Ear samples were immersed in Bouin's fixative (7 days for one-month-old mice and 30 days for mice older than 6 months) and embedded in paraffin. Sections (7 mm) were cut and mounted onto Fisher Superfrost Plus slides (Fisher Scientific, Pittsburgh, PA) and counterstained in hematoxylin/ eosin (H&E). Goblet cells, whose sole function is to secrete mucus, were identified by Mayer's Mucicarmine staining method following the protocol provided by Electron Microscopy Sciences (Catalog #26320). The stepedial artery was dissected from 6-month-old control and mutant mice, sectioned without decalcification, and stained with Alizarin red following standard procedures [21].

Evaluation of pathology of middle ears
A scoring system of -/+/++/+++ was used to assess the severity of pathology in the middle ears following a previously described method with modifications [22]. Histological analysis of the middle ears of control and mutant mice was performed using an Olympus BX51 microscope. A '-' symbol was assigned when the pathology is absent in the middle ear. A '+' symbol was assigned when pathology was very scarce in the middle ear. A '++' symbol was assigned when pathology was more prevalent, but not to the point of spanning the entire middle ear. A '+++' symbol was assigned when pathology spanned the entire middle ear. The following pathologies were evaluated by this scoring system: middle ear effusion, inflammatory cell infiltration, tissue proliferation (epithelial hyperplasia), abnormal tissue at the Eustachian tube opening in the middle ear cavity, ectopic mineralization, and clusters of goblet cells.

Scanning electron microscopy (SEM)
Middle ears from Enpp1 asj/asj and wild-type mice were dissected after transcardial perfusion with 4% paraformaldehyde (PFA) and then immersed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH = 7.2) at 4˚C overnight. Dissection was performed to expose the middle ear cavities. After three 15-minute washes with cacodylate buffer (pH = 7.2) at room temperature, samples were post-fixed with the osmium-thiocarbohydrazide-osmium-thiocarbohydrazideosmium (OTOTO) method [23], and dehydrated in increasing concentrations of ethanol at 4˚C. Samples were critical point dried with Hexamethyldisilazane (HMDS) and then air dried in fume hood. Samples were then coated with 15nm gold and analyzed in a Hitachi S-3000 scanning electron microscope (Hitachi, Tokyo, Japan) at 20 kV.

Hearing impairment in Enpp1 asj mutant mice
Hearing in the Enpp1 mutant mice and age-matched controls was evaluated by ABR threshold analysis. Initial ABR measurements of three-month-old mutant mice revealed a moderate to severe hearing loss with threshold increases of 25-35 decibels (dB) compared to the normal thresholds in the control mice for both click and all pure tone stimuli (8 kHz, 16 kHz, 32 kHz). To further characterize the hearing ability of Enpp1 asj/asj mice and to determine whether the hearing impairment is progressive with age, recurrent ABR measurements were performed in Enpp1 asj/asj , Enpp1 asj/+ , and Enpp1 +/+ littermates from 3 weeks to 30 weeks of age (Fig 1,  Table 1, Table 2). Enpp1 asj/+ and Enpp1 +/+ mice (grouped as controls) showed normal ABR thresholds even at 30 weeks of age, whereas Enpp1 asj/asj mice exhibited elevated thresholds by 6 weeks of age. Hearing impairment is not congenital as mutant mice had normal ABR thresholds at weaning age (3 weeks). Hearing loss in mutant mice progressed in two stages: average ABR thresholds increased by about 30-35 dB between 3 and 6 weeks of age, were relatively stable between 6 and 12 weeks, increased another 20-25 db between 12 and 18 weeks, and remained stable between 18 and 30 weeks. At 12 weeks of age, only one out of 11 Enpp1 asj/asj mice tested showed normal thresholds in both ears (Table 1), giving a rough estimate of hearing loss penetrance of about 90%. No head bobbing or circling behaviors were observed in the mutant mice at all ages, indicating normal vestibular function.
Otitis media with effusion in Enpp1 asj/asj mice To assess the causes of hearing impairment in the Enpp1 asj/asj mice, anatomical analysis of the middle and inner ears was performed after the completion of ABR measurements. We processed the whole bullae for histological studies, which allowed us to observe both middle and inner ear morphology. We did not observe any obvious inner ear defects in the asj mutant mice, such as hair cell and spiral ganglion cell loss, or stria vascularis degeneration. However, we observed major histopathology in the middle and outer ear, with details described below.
a. Middle ear effusion and epithelial proliferation in Enpp1 mutant mice. All ears from Enpp1 asj/asj mice that had elevated ABR thresholds (n = 11) exhibited defects in the middle ear including retracted tympanic membranes and middle ear cavities filled with effusion (Figs 2 and 3, Table 3) and a thickened epithelium with fibrous polyps (Figs 2B, 2D and 3C, Table 3). In addition, older Enpp1 asj/asj mice exhibited excessive accumulation of discharge in the external ear canal ( Fig 2H). All of the ears from the control Enpp1 +/asj and Enpp1 +/+ mice examined (n = 12) showed a transparent tympanic membrane, a clear middle ear cavity lined with a thin epithelium.
b. Increased goblet cell density and mucin secretion in Enpp1 asj/asj mice. In the epithelium lining the middle ear cavity and the Eustachian tube, mucins secreted by goblet cells build the first line of defense in protecting the host from invading pathogens; movements of the cilia then help to get rid of the pathogens, with mucins serving as a lubricant [24]. However, when chronic inflammation is present, some of the epithelial cells transdifferentiate to mucin-secreting goblet cells [24]. These additional goblet cells secrete excessive mucins, which accumulate in the middle ear cavity and lead to conductive hearing loss (otitis media). We used Mayer's mucicarmine staining to detect the density of goblet cells in the asj mutant mice. Already at two weeks of age, an excess of goblet cells was observed around the opening of the Eustachian tube and in the Eustachian tube duct of mutant mice ( Fig 4B). In the control mice, scattered goblet cells were seen around the opening of the Eustachian tube, but very few were detected along the Eustachian tube duct (Fig 4A). Similar observations were made in twomonth-old adult mice: more goblet cells were present in the epithelia lining the middle ear cavity and the Eustachian tube in mutants ( Fig 4D) than in controls (Fig 4C).
c. Impaired Eustachian tube function due to epithelia proliferation. Out of 10 ear preparations (8-30 weeks) that allowed us to observe the opening of the Eustachian tube in the middle ear cavity, 7 had amorphous tissues that could potentially block the Eustachian tube (Table 3) thereby disrupting middle ear pressure regulation and ciliary clearance of secretions. d. Impaired middle ear epithelial clearance function due to excess mucin and loss of microciliary function. Using scanning electron microscopy ( Fig 5), we assessed the integrity of the mucociliary epithelium in 1-month-old and 6-month-old wild-type and Enpp1 asj/asj mice (n = 3 each genotype). The epithelium of one-month-old mutant mice had scattered areas with high densities of goblet cells (Fig 5B), which was confirmed by Mayer's mucincarmine staining. Goblet cell density in 6-month-old mutants could not be determined because of the obscuring layer of mucin ( Fig 5D). These results indicate that excessive mucin secreted by increased numbers of goblet cells and the hindering effect of the mucin layer on ciliary function in the asj mutant mice are contributing factors to the occurrence of otitis media. Over-calcification of middle ear structures (tympanosclerosis) 6-month old control mice have transparent tympanic membrane (Fig 6A), mutant mice at the same age have retracted tympanic membrane due to the pressure of excessive discharge and white patch were observed at pars tensa of tympanic membrane in the mutant mice (Fig 6B). We dissected middle ear ossicles of 6-month-old Enpp1 asj/asj mice with littermate controls. Control mice have normal morphology of ossicles (Fig 6C). We found that stapes in the mutant mice have normal morphology and are freely removable from the round window ( Fig  6D). Although malleus an incus has relative normal morphology, these two bones are fused (Fig 6D). The wall of the stapedial artery, which passes through the ring of the stapes, is thicker  5044  3 wk  21  F  35  30  20  40  40  40  20  40   5045  3 wk  21  M  35  30  20  35  30  30  20  30   5054  3 wk  22  F  35  40  10  30  30  30  10  30   5056  3 wk  22  M  33  30  10  30  40  40  20  in Enpp1 asj mutant mice (Fig 2F) than controls (Fig 2E). Alarin red staining indicates excessive calcium deposition in the artery wall of the mutant mice ( Fig 6F). Therefore, enlarged and stiffened stapedial artery could potentially impede the movement of the stapes and lead to impaired sound transmission to the inner ear. We observed overossification of round window ridge in the mutant mice ( Fig 2B, Table 3), which may indirectly contribute to otitis media and hearing loss. We observed that the round window membrane in Enpp1 asj mutant mice ( Fig  3H) was thicker than in control mice (Fig 3G).  The effusion was confined within the middle ear cavity and did not appear to extend through the round window into the inner ear. To detect if otitis media in the asj mutant mice was caused by bacterial infection, we performed gram staining in the freshly prepared middle ear sections. We failed to detect any pathogens from the middle ear cavities of the asj mutant mice (data not shown), suggesting a non-infection origin of the otitis media in the asj mutant mice.

Discussion
Otitis media-related ear pathology of Enpp1 asj mutant mice Beginning at 6 weeks of age, an effusion starts to appear in the middle ears of asj mutant mice, and moderate epithelial proliferation is observed lining the middle ear cavity. The degree of middle ear effusion and epithelial thickening in asj mutant mice correlates with ABR thresholds. At 12 weeks of age the content of the effusion changes from serous to suppurative, with a corresponding increase in ABR thresholds. These effusions in the middle ear cavity may interfere with the normal vibration of the tympanic membrane and movement of the ossicle chain; with suppurative effusion having a much stronger effect. We observed discharge in the outer ear cavity starting around 4 months of age. All of the ears examined from seven 6~7-monthold asj mutant mice showed complete blockage of the outer ear canal by discharge, which could explain the secondary increase in ABR thresholds that occurs in asj mutant mice between 12 and 18 weeks of age. Mucins are secreted by goblet cells that lay scattered in the epithelium of digestive, respiratory, urinary, and reproductive tracts either at a basal level or at a high level upon stimulation [25]. Mucins, together with the cilia that line the epithelia, protect the host by cleaning invading pathogens. A similar mechanism is applied by the middle ear mucociliary system to clear middle ear effusions [24]. Disturbed phosphate homeostasis can cause systematic inflammation. Inflammation within the middle ear cavity triggers secretion of mucin into the middle ear cavity and transdifferentiation of more epithelial cells into goblet cells, which lead to excessive effusion accumulation in the middle ear cavity. Continued chronic middle ear inflammation leads to tissue destruction and fibrosis. Thickened middle ear epithelia or fibrosis of epithelia, especially the epithelia around the opening of the Eustachian tube in the middle ear cavity can block the Eustachian tube and facilitate effusion accumulation. Although we don't have supporting evidence, excessive discharge in the outer ear canal might be caused by inflammation in the external ear. Defective action of the cilia lining the middle ear epithelia is often associated with development of otitis media, as seen in patients with primary ciliary dyskinesia (PCD) [26]. Impaired mucociliary function and increased number of goblet cells in the middle ear cavity of the asj mutant mice, confirmed by SEM and Mayer's mucicarmine staining, indicate that the Enpp1 asj mutant mice. At 8 weeks (C), the middle ear cavity of the Enpp1 asj mutant mouse is filled with an aqueous effusion (black asterisk) and contains amorphous tissue masses (green asterisk). The middle ear epithelium is much thickened (arrow), and the otic capsule exhibits regions of ectopic mineralization (arrowhead) with adjacent fibroblastic proliferation. At 12 weeks (D), the middle ear cavity of mutant mice is filled with pus-like secretions (black asterisk), a thickened epithelium (arrows), and an amorphous tissue mass (green asterisk). E, F: Eustachian tube morphology of mutant (F) and control (E) mice at 12 weeks of age. In mutant mice, an amorphous tissue mass (green asterisk) is present near the orifice of the Eustachian tube (empty arrow), which may impede Eustachian tube function and lead to an accumulation of effusion in the middle ear cavity. G, H: Cochlear morphology and round window membranes of mutant (H) and control (G) mice. Cochlear morphology is grossly normal in Enpp1 asj mutant mice compared with controls; however, the round window membrane (black arrow) of mutant mice is thicker. MEC: middle ear cavity. Scale bars: A, B, C, D, E, F = 200μm, G, H = 100μm.
doi:10.1371/journal.pone.0168159.g003 Table 3. Histological assessment of the middle ears of the Enpp1 mutant and control mice. MEE: Middle Ear Effusion; EH: Epithelial Hyperplasia; IC: Inflammatory Cells; ETO: Eustahian Tube Opening; EM: Ectopic Mineralization; CGC: Cluster of Goblet Cells. Age  Genotype  MEE  EH  IC  ETO  EM  CGC   4012L  1wk  Control  -+  -N/A  --4017  1wk  Control  ---N/A  --4015L 1wk  Mineralization disorders and conductive hearing loss in Enpp1 asj/asj mice Mineralization disorders due to abnormal phosphate levels have been associated with inflammation in several different diseases, including chronic kidney disease [27,28] and  Hearing Loss in Enpp1 Mutant Mice caridiovascular disease [29]. Enpp1 asj adds to a growing list of mouse mutations causing phosphate homeostasis disorders that have associated conductive hearing loss, including mutations of the Ank, Phex, Rpl38, and Fgf23 genes. Ank (progressive ankylosis) encodes a multiple-pass transmembrane protein that regulates pyrophosphate levels, and Ank mutant mice were reported to have middle ear ossicle fusions and associated conductive hearing loss [30]. ANK mutations in human patients also have been reported with associated with conductive hearing loss [31]. PHEX (X-linked phosphate regulating endopeptidase) is an enzyme that is involved in regulating the balance of phosphate in the body. PHEX mutations are associated with Xlinked hypophosphatemic rickets in human patients, with hearing loss as one of the symptoms [32][33][34]. Mice with Phex mutations exhibit hypophospatemia-related abnormalities and hearing impairment [35], which recently was shown to be associated with middle ear effusion and ciliary defects [8]. Rpl38 (ribosomal protein L38) is not known to be directly involved in phosphate regulation; however, elevated organic phosphate levels in Rpl38 Ts mutant mice [4] suggest a potential but unknown function of Rpl38 in phosphate homeostasis. Rpl38 Ts mutant mice have conductive hearing loss caused by ectopic ossification and cholesterol crystal deposition in the middle ear cavity, enlarged Eustachian tube, and chronic inflammation with effusion [4]. FGF23 (fibroblast growth factor 23) is a circulating hormone that controls phosphate and calcium homeostasis and bone mineralization. Abnormal serum levels of FGF23 lead to systemic pathologies in humans, including renal phosphate wasting diseases and hyperphosphatemia. FGF23-deficient mice show a mixed hearing loss and middle ear malformations [36], and changes in circulating FGF23 have been observed in humans and mice with ENPP1 and PHEX deficiencies [16,37]. ENPP1 produces inorganic pyrophosphate (PPi), an inhibitor of mineralization. Because PPi levels are markedly reduced in Enpp1 asj/asj mice [13], it is not surprising to see ectopic mineralization and calcification of middle ear tissues, as was observed in other soft tissues of these mutant mice [13]. Calcification of soft tissues within the middle ear cavity could contribute in various ways to the conductive hearing loss of Enpp1 asj/asj mice. The stapedial artery is an embryonic artery that disappears at the early embryo stage in humans but is conserved in mice through adult ages [38]. We observed a thickened stapedial artery wall in Enpp1 asj mutant  The tympanic membrane appears to be transparent and malleus (blue asterisk) is clearly visible. In the age matched mutant mice (B), surrounding bone must be removed to expose the tympanic membrane, most of which is completely covered by discharge. White patches (inside black dashed lines) were observed on the tympanic membrane of the mutant mice (B). The ossicle bones of the age matched control and mutant mice appear to have similar morphology (C-D), but malleus and incus are fused in the mutant mice (D, red arrow). M: Malleus; I: Incus; S: Stapes. Alizarin red staining reveals extensive mineralization in the stapedial artery wall (green asterisk) in Enpp1 as/asjj mice (F), but not in the control mice (E).

Mouse ID
doi:10.1371/journal.pone.0168159.g006 mice, which is likely due to calcification as has been observed in other soft tissues. The thicker arterial wall is not likely to affect transport function, but because the artery passes through the ring of the stapes, its thickened wall could potentially impede the vibration of the stapes and contribute to the conductive hearing loss of mutant mice.
Abnormal mineralization can also cause otosclerosis and tympanosclerosis, two conditions that commonly lead to hearing loss in human patients. Disruption of bone homeostasis of the otic capsule can lead to otosclerosis [39], and the most common feature of otosclerosis is stapes fixation. Although we observed ectopic mineralization and bone deposition in the otic capsule of asj mutant mice starting from around 6 weeks of age, stapes fixation is absent from Enpp1 mutant mice. Instead, we observed white patch on the tympanic membrane, malleus and incus fusion, which are typical symptoms of tympanosclerosis. We observed a thickened round window membrane in Enpp1 asj/asj mice at 8 weeks of age, which may increase the rigidity of the membrane and impede proper cochlear fluid movement and hair cell stimulation [40]. Therefore, Enpp1 asj mice can serve as a model for studying tympanosclerosis.
Decreased PPi levels in Enpp1 asj/asj mice also lead to otitis media, perhaps the most important factor contributing to the hearing loss, but the underlying mechanism of pathology is uncertain. ENPP1 deficiency is known to cause elevated serum levels of FGF23 in Enpp1 mutant mice [16], and excess FGF23 secreted in the middle ear may trigger mucoperiosteum proliferation, which may contribute to the development of otitis media. In support of this possibility, mucoperiosteum proliferation in Enpp1 asj/asj mice is remarkably enhanced around the regions of the otic capsule that exhibit ectopic mineralization. Enpp1 asj mice provide a tool to unravel the underlying mechanism of the development of otitis media that is associated with abnormal phosphate homeostasis. This is the first report of hearing loss and ear pathology associated with a mutation of the mouse Enpp1 gene. The conductive hearing loss of Enpp1 asj mutant mice provides a new animal model for studying otitis media and tympanosclerosis related to mineralization defects. It also provides a specific model for understanding the hearing loss recently reported to be a clinical feature associated with human ENPP1 mutations [17,18].