| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* URPhyM, Laboratory of Physiology and Pharmacology, University of Namur, Namur, Belgium;
Department of Biochemistry and Medical Genetics, The University of Manitoba, Winnipeg, Manitoba, Canada; and
Halozyme Therapeutics, San Diego, California, USA
1Correspondence: URPhyM, Laboratory of Physiology and Pharmacology, University of Namur, 61 rue de Bruxelles, 5000 Namur, Belgium. E-mail: bruno.flamion{at}fundp.ac.be
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: hyaluronan • hyaluronidase • sinusoidal endothelial cells • hemolysis • frontonasal suture
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
.
HA is produced at the plasma membrane by three different mammalian HA synthases: HAS1, HAS2, and HAS3 (3)
. Once outside the cell, HA can distribute into the ECM or remain attached to the cell surface through various receptors. HA may also reside within cells, either following receptor-mediated or fluid-phase endocytosis or during inflammatory processes (4)
. HAS2 deficiency in mice results in embryonic lethality at E9.5 linked to a defect in cardiac morphogenesis (5)
; HA is almost absent in HAS2 knockout (KO) tissues.
On the degradation side, somewhat surprisingly, this simple GAG structure seems to require multiple hyaluronidases (HYALs). Hyal1, Hyal2, Hyal3, Hyal4, PH-20, and Hyalp1 genes have been identified in mammals; they are clustered on two different chromosomes (6)
. In the mouse, a seventh gene has recently been discovered and is now referred to as Hyal5 (7)
. However, the roles of each HYAL in HA turnover either individually or coordinately remain poorly understood. The most active HYALs are HYAL1 and PH-20. The latter is a glycosylphosphatidylinositol (GPI) -anchored cell surface protein that resides almost exclusively in the sperm, has acid and neutral pH optima, and is involved in fertilization (8)
. HYAL1 is a secreted protein present in many tissues and plasma; it shows little activity at pH > 4.0 (9)
. HYAL1 deficiency has been reported to cause a lysosomal storage disorder in humans called mucopolysaccharidosis IX (10
, 11)
. The only patient described so far exhibited a mild phenotype with periarticular soft tissue masses, short stature, and acetabular erosions. Lysosomes of macrophages and fibroblasts in affected tissues were filled with HA. Plasma hyaluronidase activity was deficient and plasma HA was grossly elevated. In the mouse, the primary manifestation of HYAL1 deficiency is in the joint where accumulation of HA in chondrocytes is accompanied by the progressive loss of proteoglycans (12)
. The serum HA levels were not elevated in this mouse model, and no HA accumulation was detected outside of the skeleton (12)
. Overall, HYAL1 is proposed to be a lysosomal enzyme, which might be internalized from the extracellular medium (13)
.
HYAL2, which like PH-20 is also GPI-anchored (14)
, has a much weaker hyaluronidase activity than HYAL1 or PH-20; titration of the enzymes shows secreted HYAL2 to require an acidic pH for activity and to be 
400-fold less potent than PH-20 (15)
. Fragments of HA 20 kDa in mass may accumulate during HYAL2 functioning (13
, 15)
. The physiological roles of the other HYALs remain uncertain. The most accepted model of HA degradation has been proposed by Stern (16)
. The author hypothesized that HA catabolism begins at the cell membrane, where it binds to its major receptor CD44 and becomes internalized in caveolae. HYAL2 would then cleave HA to intermediate-size fragments that are endocytosed and further degraded in lysosomes by HYAL1, β-N-acetylglucosaminidase, and β-glucuronidase. However, the overall evidence for a two-step model of HA degradation is scarce.
HYAL2 has become of major interest because there is a growing body of data indicating that the length of the HA molecule affects its biological functions (17
, 18)
, and HYAL2 is proposed to be one of the main actors regulating the size of HA, especially during inflammation (18
, 19)
. HYAL2 has also been revealed as the cellular receptor for some oncogenic ovine retroviruses (14
, 15)
. Nevertheless, the involvement of HYAL2 in HA catabolism and homeostasis remains unclear, and little is known about its actual functions. To clarify its roles, we wanted to generate a mouse model that is deficient for this enzyme. Since unsuccessful attempts had apparently been made to produce viable HYAL2 KO mice (19)
, we decided to generate conditional KO mice using the Cre-loxP system. Viable HYAL2-deficient mice with skeletal and hematological defects were obtained, demonstrating that HYAL2 has a physiological function during development and later in life. This article contains a description of the generation and observable phenotype of Hyal2–/– mice and suggests that a new form of mucopolysaccharidosis IX is yet to be defined in humans.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Design of the replacement vector
A fragment of 129/Ola mouse genomic DNA containing Hyal2 was obtained from a cosmid library (clone ID MPMGc121P09180Q2, RZPD, Berlin, Germany), digested with restriction endonucleases NotI and SalI, and cloned into pBlueScript SK+ (Stratagene, La Jolla, CA, USA). Through successive subcloning steps, a loxP site was inserted in a XmnI restriction site present in the first intron of Hyal2, and a neomycin resistance cassette (PGK-neo) flanked with two loxP sites was inserted in an ApaI restriction site downstream of the gene as a selection marker. This resulted in floxing of the complete coding sequence of Hyal2 (Hyal23lox) since exon 1 is noncoding. The disappearance of the XmnI and ApaI restriction sites used to introduce the loxP site and the selection marker was used for identification of the targeted allele in subsequent Southern blot analyses. The replacement vector was linearized and electroporated into E14 (129/Ola) embryonic stem (ES) cells that were cultured on G418-resistant mouse embryonic fibroblasts in Glasgow minimum essential medium (GMEM; Invitrogen, Carlsbad, CA, USA) supplemented with 2 mM L-glutamine, 1 mM sodium pyruvate, nonessential amino acids, 55 nM β-mercaptoethanol, 10% FBS (Hyclone, Logan, UT, USA), 1000 U/ml LIF (Chemicon, Temecula, CA, USA), and 250 µg/ml G418. Four hundred resistant colonies were picked up and screened for homologous recombination by polymerase chain reaction (PCR) using a sense primer hybridizing to the PGK-Neo cassette (5'-GCGGGGATCTCATGCTGGA-3') and an antisense primer hybridizing downstream of the SalI restriction site (5'-GGGACACTCAAAGACAGCTG-3'). Correctly targeted clones were identified by standard Southern blot analysis using 5' and 3' probes external to the regions of homology and following either XmnI or ApaI digestion. The 5'-probe was PCR-amplified using 5'-CCTCCTCTAAGGGAAAGGTTGC-3' and 5'-AATGGCTGACGCTCAAATGC-3' as primers. The 3'-probe was PCR-amplified using 5'-TGAATGCCTAAGGGGTTCGG-3' and 5'-TCTGTGCCTGCTGTGCCTA ATG-3' as primers. Eight correctly targeted clones were obtained.
Generation of Hyal2+/3lox and Hyal2+/– mice
Cells from a targeted clone were aggregated with CD1 premorula-stage embryos to generate male chimeras that subsequently transmitted the mutation through the germline (20)
. The offspring carrying the Hyal23lox allele were bred with PGK-Cre transgenic mice in order to achieve complete recombination to generate the Hyal2-deleted allele. Genotypes were determined using Southern blot and PCR analysis. Southern blot analysis was carried out using XmnI digested tail DNA and the 5'-probe mentioned above. For PCR analysis, DNA from ear-punches was amplified with specific sense primers for the wild-type allele (5'-GGCCTGACTCTACTCCCTG-3') and the Hyal2-null allele: (5'-CGGTGCCCGAGACTAAGTC-3'), and with a common antisense primer (5'-AGGAACATCAGGGAAGATCAT-3').
Reverse transcriptase-PCR (RT-PCR) analysis
Total RNA was extracted from rotor-stator homogenized mouse tissues using Illustra RNAspin Minikit (GE Healthcare, Chalfont St. Giles, UK) and reverse transcribed using MMLV-reverse transcriptase. Forty nanograms of cDNA was amplified using specific primers for Hyal2 (5'-GGCTCTCTTTCCCTCTGTGTA-3' and 5'-GGGAACCAGCAGCTGAGTTA-3'), Rassf1 (5'-AGCATCTGCATGTTCTATCACG-3' and 5'-GTGTTCCTCTTCTTCCCGCT-3'), Tusc2 (5'-AGCAGTCTCTGGTGCGGTCT-3' and 5'-TCCTGGGAGTTTCTTGCTGG-3'), Nat6 (5'-CTGAGAACTTAGGAGGAGGTG-3' and 5'-ATCTCAGGGTCAACTCTGGC-3'), and Ifrd2 (5'-GAGGAGCTGTTCCGTAGCCT-3' and 5'-ATGGCTGATGTGGGTACCAG-3'), and run on a 1.5% agarose gel in order to determine their level of expression. Gapdh was amplified as an endogenous control (5'-ATCTTCCAGGAGCGAGACCC-3', and 5'-CATGAGCCCTTCCACAATGC-3').
Real-time PCR analysis
For Hyal1 and Hyal3 quantitation, total RNA was extracted as described above. Ten nanograms of reverse transcribed RNA was amplified in triplicate in an Applied Biosystems 7300 device using Power SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA, USA) with specific primers for Hyal1 (5'-TGCCCGTAATGCCCTACGT-3' and 5'-GCTGTGCTCCAGTTCCTCCA-3'), Hyal3 (5'-CCGGAGCTCTGGGAGATTC-3' and 5'-GCGGCACTCACTCCAATAGTC-3'), and Actb (5'-ACGGCCAGGTCATCACTATTG-3' and 5'-CACAGGATTCCATACCCAAGAAG-3'). The following profile was used: 2 min at 50°C, 10 min at 95°C, then 40 cycles of 15 s at 95°C and 1 min at 60°C. Relative quantification was performed using Actb for normalization. For Hyal2 expression profiling, pooled cDNAs from different mouse tissues were purchased from Clontech (Mountain View, CA, USA) (Mouse MTC Panel I). One nanogram of each cDNA was amplified under the same conditions but with primers specific for Hyal2 (5'-CGAGGACTCACGGGACTGA-3' and 5'-GGCACTCTCACCGATGGTAGA-3'), and, as normalizers, Actb (see above), Tbp (5'-CAGTTACAGGTGGCAGCATGA-3' and 5'-TAGTGCTGCAGGGTGATTTCAG-3'), and Rpl37 (5'-GGTCCCCATCGTGTTTGC-3' and 5'-TCAGCTCATCACTCACCCACTT-3'). The expression profile of each normalizer was analyzed, and outliers were removed. The level of Hyal2 mRNA was quantified in relation to each normalizer and the mean value plotted on a graph.
HA zymography for hyaluronidase activity
Rooster comb HA (Sigma, St. Louis, MO, USA) at a final concentration of 0.17 mg/ml was added to the 8% polyacrylamide solution before polymerization of gels. Fresh plasma samples were dissolved in Laemmli buffer at room temperature in the absence of reducing agent and separated by SDS-PAGE on ice. Gels were washed in 3% Triton-X100 for 2 h at room temperature and incubated overnight at 39°C in 0.1 M formate, pH 3.7 containing 0.1 M NaCl. Gels were then treated for 2 h at 39°C with 0.1 mg/ml pronase in 20 mM Tris-HCl, pH 8.0, and stained for at least 24 h with 0.015% Stains-All (Sigma) in 50% formamide. The gels were briefly destained in water and photographed with a digital camera on a transilluminator.
HA levels and quantitative hyaluronidase assay in murine plasma
The levels of hyaluronidase activity were determined using a modified biotinylated substrate assay (21)
. Briefly, 100 µg/ml bacterial HA (Lifecore, Chaska, MN, USA; molecular mass 1100 kDa) coupled with biotin hydrazide at a ratio of 1:100 disaccharides was coated on Thermo-Immulon 4HBX 96-well flat-bottom plates (Thermo Fisher Scientific, Waltham, MA, USA) in Na-bicarbonate buffer, pH 9.6. BSA-blocked wells were incubated for 120 min at 37°C with murine plasma samples diluted 1:50 in 0.1 M formate, pH 3.7, with 50 mM NaCl and 5 mg/ml BSA. Hyaluronidase-dependent liberation of biotinylated HA was detected with streptavidin horseradish peroxidase (HRP) and a tetramethylbenzidine (TMB) substrate calibrated against a standard curve of immunoaffinity-purified recombinant human HYAL1 (EC50=0.68 ng/ml at pH 3.7 with 120 min incubation). HA levels were determined using a competitive ELISA assay. Plasma samples were incubated for 60 min with 6 ng/ml biotinylated HA binding protein (bHABP, Seikagaku Corp., Tokyo, Japan) in PBS with 0.5 M NaCl, 0.05% Tween, and 0.5 mg/ml BSA prior to transfer to an HA-coated plate for an additional 60 min incubation. Free bHABP available to bind to the HA-coated plate was detected using streptavidin HRP and compared to a 2–128 ng/ml standard curve of 1100 kDa HA (Lifecore). In some experiments, samples were predigested with proteinase-K or treated with recombinant human PH-20 hyaluronidase followed by heat inactivation prior to the addition of bHABP to confirm specificity.
Skeleton preparation and morphometric measurements
Either whole skinned and eviscerated mice or skinned heads alone were prepared. Samples were fixed overnight in 95% ethanol, then stained for cartilage with 0.015% Alcian blue (Sigma) in 75% ethanol/20% acetic acid for 24 h. Samples were washed 3 times in 95% ethanol, the last wash being carried on overnight, then cleared in 2% KOH for 24 h, stained for bone with 0.005% alizarin red in 1% KOH overnight, cleared in 1% KOH/20% glycerol for 48 h, and finally stored in 50% ethanol/50% glycerol. Nose length and interorbitary space were measured using a digital caliper.
Histological analysis
Samples were dissected, fixed for 24 h in 4% buffered paraformaldehyde, and decalcified overnight in RDO (Eurobio, Courtaboeuf, France). They were then embedded in paraffin and cut into 6-µm sections that were stained with hematoxylin or hemalun, erythrosin and safran (HES). Coverslips were attached with dibutyl phthalate xylene (DPX), and slides were observed under a bright-field microscope.
Blood analyses
Heparinized blood was collected by retro-orbital bleeding. Cell count, hemoglobin concentration, and hematocrit were measured using a Bayer Advia 120 automated hematology analyzer (Bayer, Leverkusen, Germany). Urea, creatinine, LDH, and glucose measurements were performed using standard procedures.
Blood smears
Heparinized blood was collected by retro-orbital bleeding. Three microliters was spread out onto a microscope slide and air-dried. Slides were fixed and stained with May-Grunwald and Giemsa stains, then incubated 1 min in 1% boric acid and left to dry. Coverslips were attached with DPX, and slides were observed under a bright-field microscope.
HA detection in mouse tissues
Samples were dissected, fixed for 24 h in 4% buffered paraformaldehyde, and embedded in paraffin. Six-micrometer sections were collected on slides, rehydrated, treated with 0.1 M glycine and 3% H2O2 to eliminate endogenous peroxidase activity, then incubated with 5 µg/ml bHABP in PBS containing 0.1% BSA and 0.02% Triton X-100. Sections were incubated with streptavidin-HRP (Dako, Glostrup, Denmark), stained with AEC+ (Dako), and counterstained with hemalun. Control experiments in which bHABP was omitted were performed. Coverslips were attached with Glycergel (Dako), and slides were observed under a bright-field microscope.
Iron detection in mouse tissues
Samples were dissected, fixed for 24 h in 4% buffered paraformaldehyde or Bouin’s fixative, embedded in paraffin, and cut into 6-µm sections. Sections on slides were deparaffinized and stained using Perl’s method. Briefly, slides were rinsed in distilled water, then incubated for 15 min in 5% potassium hexacyanoferrate II and for 30 min in 2.5% potassium hexacyanoferrate II and 5% HCl, rinsed again in distilled water, and counterstained with HES. Coverslips were attached with DPX, and slides were observed under a bright-field microscope.
Statistical analyses
Results are expressed as means ± SE. Statistical significance was set at P < 0.05 and was determined using t tests for single comparisons and ANOVA for multiple comparisons. A 
2 test was used to analyze the proportion of different genotypes in heterozygous intercrosses and a log-rank test to evaluate the statistical significance of differences in survival curves.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, we decided to generate conditional KO mice in order to study the effect of inactivation of HYAL2 on different processes and tissues. A Cre-loxP system with a "three loxP sites" strategy was selected. A replacement vector was constructed containing a mutated Hyal2 allele, flanked by a loxP site on one side and a floxed neomycin cassette (PGK-Neo) on the other side (Fig. 1A
). This allele is referred to as Hyal23lox. Hyal2 is a small gene, 3.65 kb long, composed of 4 exons, the first of which is noncoding. To ensure Cre recombination would generate a null allele, the last three exons of Hyal2, i.e., the full coding sequence, were floxed, precluding the production of any functional protein fragment.
|
The replacement vector was electroporated into E14 mouse ES cells, and several cells underwent homologous recombination, as subsequently detected by PCR and Southern blot analysis of DNA from colonies using two different probes with two different restriction profiles (not shown). The clones that were selected based on this screening strategy were confirmed to have only one recombination event by another Southern blot analysis, using a probe hybridizing to the PGK-Neo cassette (not shown).
The cells from a correctly targeted ES cell colony were then aggregated with CD1 embryos to generate male chimeras that were able to transmit the mutated allele through their germline. Once Hyal2+/3lox mice had been obtained, they were bred to transgenic mice expressing Cre under control of the PGK promoter. This led to the excision of the full coding sequence of Hyal2, resulting in the creation of a Hyal2 null allele (Fig. 1B, C
).
Hyal2–/– mice are viable and fertile
Hyal2+/– mice were bred together to determine whether viable Hyal2–/– animals could be obtained. The genotype of the first litters was determined by PCR analysis and verified by Southern blot analysis (Fig. 1D
). Both Hyal2–/– and Hyal2+/+ offspring were identified in these litters. Subsequent litters were genotyped by PCR analysis alone. RT-PCR analysis of liver and kidney total mRNA confirmed that HYAL2 was not expressed in Hyal2–/– mice and that the expression of the genes closest to Hyal2 did not appear to be decreased (Fig. 1E
). The levels of expression of Hyal1 and Hyal3 were determined more precisely using real-time RT-PCR (see
Fig. 6C
). Hyal2–/– mice were viable and fertile, though a significant deficit in Hyal2–/– animals at weaning among the Hyal2+/– intercrosses was observed (18% instead of the expected 25%; P<0.02, n=309), and survival appeared slightly lower in Hyal2–/– than in Hyal2+/– and Hyal2+/+ mice over an 18-month observation period (Fig. 2
).
|
|
|
|
|
Hyal2 mRNA is ubiquitous
Quantitative PCR was performed on first-strand cDNA from various wild-type tissues and embryonic stages. Hyal2 was expressed in all tested tissues (Fig. 3
). The lung, liver, and kidney had high expression; the brain had the lowest levels.
Hyal2–/– mice display craniofacial and cervical vertebrae abnormalities
Externally, Hyal2–/– mice appear healthy but can generally be recognized among litters by their characteristic shortened nose as early as 3–4 wk (Fig. 4A
). This anomaly persists into old age. Chondroskeletal preparations of adult mice revealed abnormal craniofacial bones with an extra diamond-shaped bony structure inserted between frontal and nasal bones, deforming the anterior part of the metopic suture and the frontonasal junction (Fig. 4B
). This bony structure was already present at birth but was not detected in E17.5 embryos (not shown). Morphometric measurements performed on skulls obtained from 2- to 3-month-old male (n=10) and female (n=8) mice confirmed the shortening of the nose and revealed a significantly widened interorbitary space in Hyal2–/– males compared to wild-type littermates (Fig. 4C
). Hyal2+/– mice were similar to wild-type mice (not shown).
In addition, the first two cervical vertebrae of Hyal2–/– mice exhibited an abnormal shape, with enlarged bodies and apophyses (Fig. 4D
). The odontoid process was often fused with C1 instead of C2. No obvious anomaly could be detected on any other skeletal element.
Histological analysis reveals a defect in the junction between C1 and C2 in Hyal2–/– mice
On decalcified, paraffin-embedded skull sections, the abnormal junction between frontal and nasal bones was clearly visible, but no other underlying defect could be observed (Fig. 5
). Analysis of cervical spine sections revealed a defect in the C1–C2 junction. The odontoid process was fused to C1 instead of C2 (Fig. 5)
. No other difference could be observed between Hyal2–/– and wild-type mice in these sections. There was no difference in cortical bone thickness.
Hyal2–/– mice display grossly elevated plasma HA levels and moderately increased plasma hyaluronidase activity
Plasma HA levels were significantly increased in Hyal2–/– mice (Fig. 6A
). Mean plasma levels were 601 ± 112 and 6490 ± 1100 ng/ml in Hyal2+/+ and Hyal2–/– mice respectively (P<0.01), which is a 10-fold increase. Similar values were obtained following proteinase-K digestion and heat inactivation of plasma samples prior to addition of bHABP, ruling out matrix effects from changes in plasma HA binding proteins. Predigestion of plasma samples with excess recombinant human PH-20 followed by heat inactivation prior to bHABP addition eliminated HA detection, demonstrating specificity of the bHABP. Using both HA zymography and a biotinylated HA substrate microtiter assay, a 2-fold increase in plasma hyaluronidase activity in Hyal2–/– mice compared to their control littermates was observed (Fig. 6A, B
). Using an immunoaffinity-purified recombinant human HYAL1 standard assayed at pH 3.7, mouse plasma hyaluronidase levels were estimated to be 21 ng/ml in Hyal2+/+ mice and 40 ng/ml in Hyal2–/– mice. This activity in murine plasma is undoubtedly due to the main plasma hyaluronidase, i.e., HYAL1, as shown by comparison with zymographic analyses of the plasma of Hyal1–/– mice performed under identical conditions (Fig. 6B
) and the absence of enzyme activity in Hyal1–/– mice by biotinylated microtiter assay (not shown).
Tissue expression of Hyal1 and Hyal3
The amounts of Hyal1 and Hyal3 mRNA were evaluated using real-time RT-PCR in liver and kidney extracts since Hyal2 mRNA is abundant in these tissues. Hyal1 was significantly up-regulated in the kidney (P<0.001) but not in the liver; the expression of Hyal3 remained very low in all cases and did not change significantly in Hyal2–/– mice (Fig. 6C
).
Tissue HA detection
HA was identified in various tissues from 11- to 12-wk-old male and female mice using a specific HA binding protein. No differences in tissue HA staining were observed in the kidney, lung, spleen, heart, skin, or bones between Hyal2+/+ and Hyal2–/– mice (not shown). However, a significant accumulation of HA was detected in Hyal2–/– livers but only at specific locations that appear closely linked to liver sinusoidal endothelial cells, and perhaps Kupffer cells, rather than hepatocytes (Fig. 7
).
|
Hyal2–/– mice display altered blood parameters
Complete hematological counts were performed on the blood of Hyal2–/– and wild-type mice. The results are shown in Table 1
. A mild anemia with high reticulocyte counts was observed in Hyal2–/– mice. The platelet counts were also decreased but no sign of altered red blood cell morphology was noted in blood smears apart from polychromatism (Fig. 8
), suggesting that thrombocytopenia was not due to microangiopathy. Bone marrow histology looked normal (data not shown). Plasma lactate dehydrogenase was elevated, which is compatible with a hemolytic anemia. Renal function appeared unaltered, as shown by urea and creatinine levels within the normal range. Because increased plasma hyaluronidase has been associated with diabetes (22
, 23)
, glycemia was also measured in Hyal2–/– and control mice; no difference was observed.
|
|
Iron deposits in Hyal2–/– mice
Brown pigments could be observed on HES-stained Hyal2–/– tissue sections (not shown). Because these pigments were suspected to be iron, Perl’s staining was performed that confirmed the presence of iron deposits in the liver, the kidney, and the spleen (Fig. 9
). In the kidney, these deposits were prominent and exclusively located in proximal tubules, sometimes very close to the space of Bowman. This observation, when linked to reduced erythrocyte and elevated reticulocyte counts, as well as elevated plasma lactate dehydrogenase, strongly supports the hypothesis that Hyal2–/– mice suffer from chronic intravascular hemolysis.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In a first attempt to characterize HYAL2 deficiency, these mice were bred to transgenic mice that strongly express the Cre recombinase (PGK-Cre mice). This resulted in the deletion of the full coding sequence of Hyal2 in the targeted locus. By breeding mice heterozygous for this null allele, we obtained mice deficient in HYAL2, i.e., Hyal2–/– mice. These mice were viable, fertile, and apparently healthy, though a mild deficit in KO offspring from heterozygous intercrosses was observed. The precise time at which some of the KO pups die is not known, since the mice were genotyped at weaning, i.e., at 3 wk of age. An apparent excess of premature deaths was also observed among KO mice up to 18 months of age. Further study will be needed to determine whether Hyal2–/– mice are more susceptible to specific diseases.
The absence of Hyal2 expression in KO mice was confirmed by RT-PCR analysis in liver and kidney. Examination of the genes closest to Hyal2, i.e., Hyal1, Hyal3, Rassf1, Tusc2, Nat6, and Ifrd2, showed that they were still expressed. We can thus assume with confidence that the observed phenotype is actually due to the absence of HYAL2.
Hyal2 KO mice were checked for gross external abnormalities. The only difference in appearance was the characteristic face of KO mice, exhibiting a shortened nose and wide-set eyes. This was due to the presence of an additional, diamond-shaped bone between the frontal and nasal bones. The presence of a narrow interfrontal bone has been reported previously in some laboratory strains such as C57BL/Gr and in samples of wild mice, but this minor skeletal variant does not disrupt the frontonasal process (24
, 25)
. The latter is formed by intramembranous osteogenesis of cells originating from the neural crest and migrating through specific pathways to reach their destination where they start to differentiate and condense into bone (26)
. Disruption in several genes, including Msx2 and Twist, as well as the combined mutation of Bgn/Dcn, result in a lack of ossification of the frontal bones and impaired posterior frontal sutural fusion, sometimes with large skull vault foramen (27
, 28)
. However, none of these mutations leads to the abnormalities observed in Hyal2–/– mice. We speculate that HYAL2 could be involved in regulating the migration of the neural crest cells to a specific frontonasal location without affecting other neural crest-derived features such as skin pigmentation and formation of the palate and mandible. The role of HYAL2 in the formation of the frontonasal suture could also be more indirect, affecting intramembranous osteogenesis itself, although in that case, other craniofacial bone defects would also be expected as in the Msx2 and Twist mutants (27)
.
In addition, Hyal2–/– mice presented with a defective separation between the first two cervical vertebrae, the odontoid process being merged with the atlas instead of the axis (see Fig. 5
). Intervertebral discs appeared to be histologically normal. Several congenital anomalies of the odontoid process have been described in humans, including its detachment from the axis and the persistence of the ossiculum terminale, i.e., failure of fusion of the terminal ossicle to the remainder of the odontoid process (29
, 30)
. To our knowledge, none of these anomalies has been linked to a specific mutation so far. Disconnection of the odontoid process from the axis may lead to spinal compression. We cannot exclude the possibility that the decreased survival of HYAL2-deficient mice is due to cervical myelopathy-induced sudden death, even though no obvious neurological signs were detected during their lifetime. Interestingly, abnormalities of the odontoid and craniofacial dysmorphism can be found in several human and animal mucopolysaccharidoses as well as in Down’s syndrome (e.g., 31
32
33
).
Quantitative RT-PCR analysis of adult mouse tissue mRNAs confirmed that Hyal2 expression is ubiquitous and closely mirrors prior observations in human tissues showing abundant Hyal2 mRNA in the heart, placenta, lung, liver, skeletal muscle and kidney coupled with very low levels in the brain (34
, 35)
. HYAL2 deficiency may thus produce alterations in various cell and organ functions throughout life and may modify the animal’s response to numerous stimuli. As a preliminary analysis of potential steady-state alterations in the Hyal2–/– mouse, we measured plasma HA and hyaluronidase levels, as well as hematological parameters. Plasma hyaluronidase activity was moderately elevated in KO animals, but most surprisingly, plasma HA levels reached 10-fold the mean normal values. Such changes were unexpected given that the enzymatic activity of HYAL2 is reported to be low and incomplete (13
, 15
, 34)
. Our study clearly demonstrates that HYAL2 has a physiological activity that helps maintain relatively low concentrations of plasma HA. The details of this activity remain to be sorted out. Previous examples of extremely high levels of plasma HA have been limited to a few human cases of neuroblastoma, mesothelioma, hemangioendothelioma, Wilms tumors, and liver failure (36
, 37)
but have not been described in mice to our knowledge. The origin of high blood HA in the above patients has not been elucidated. A single, unexplained human case of skin thickening through massive accumulation of HA also displayed very high plasma HA concentrations (38)
; however, Hyal2–/– mice have a normal skin appearance and HA staining.
Among several tissues probed with a HABP, none displayed any gross accumulation of HA in Hyal2–/– mice. The only exception was the liver, in which HA appeared to accumulate in sinusoidal cells, suggestive of endothelial cells and/or Kupffer cells, or at least in close relation to these cells. It is generally accepted that liver endothelial cells are the main site of clearance of circulating HA (1)
; for instance, Fraser et al. (39)
reported that 90% of intravenously injected radioactive HA is taken up by sinusoidal endothelial cells but none by Kupffer cells. The entry of HA into sinusoidal endothelial cells is mediated by a specific scavenger receptor called HARE (40)
; a monoclonal antibody to HARE completely blocks 125I-HA uptake by the isolated perfused rat liver (41)
. The very high plasma HA concentrations in Hyal2–/– mice may simply overload the sinusoidal clearance mechanism, leading to the accumulation of HA. Alternatively, we cannot exclude the possibility that a lack of HYAL2 within sinusoidal cells leads to both cellular HA accumulation and defects in receptor-mediated HA uptake, as shown e.g., in experimental damage to endothelial cells, ischemia-reperfusion injury, and liver transplantation (42
, 43)
. Defective HA clearance by Hyal2–/– liver sinusoidal endothelial cells could thus lead to high plasma HA concentrations and Kupffer cell HA overload.
In addition, a 2-fold increase in plasma hyaluronidase activity, which can be attributed to HYAL1, and increased amounts of Hyal1 transcripts in the kidney, but not in the liver, were detected in Hyal2–/– mice. This cannot explain the high plasma HA levels. On the contrary, HYAL1 deficiency in humans, i.e., MPS IX, is linked to elevated plasma HA concentrations, up to 2000 ng/ml, i.e., 40 to 100 times the mean normal value of 23.7 ± 14.6 ng/ml (10)
. Hyal1–/– mice, a model of MPS IX, have normal or barely elevated plasma HA concentrations (12)
. However, the nearly 30-fold difference in plasma HA levels observed between wild-type Hyal1+/+ mice and human plasma may mask the changes in plasma seen in MPS IX. In comparison, Hyal2–/– mice have a 10-fold increase in plasma HA, reaching concentrations of 6500 ng/ml.
In summary, it appears that both HYAL1 and HYAL2, but especially the latter, are needed to maintain plasma concentrations of HA at a normal level, at least in mice. These findings are consistent with the model proposed by Robert Stern (16)
, in which HYAL2 functions to initiate, and HYAL1 to continue, HA degradation. A deficiency in HYAL2 would be expected to lead to extracellular accumulation of HA, as is seen in our study, where the main site of HA accumulation is the blood. The absence of HA deposits within cells in several organs examined, apart perhaps from sinusoidal cells, suggests that any role HYAL2 may have inside of the cell is minimal or redundant with the function of other hyaluronidases.
Another remarkable feature of HYAL2 deficiency is a mild anemia with an increased number of circulating erythrocyte precursors, high plasma LDH levels, and iron deposits in the liver, spleen, and kidney. Iron accumulation in the renal proximal tubules was particularly striking, suggesting chronic intravascular hemolysis. Simultaneously, a moderate thrombocytopenia was noted, without any sign of microangiopathy or renal disease (apart from the tubular iron deposits). Further studies will be needed to determine whether the defects in erythrocyte and platelet counts are intrinsic to these elements, maybe through a direct lack of HYAL2, or whether they are secondary to the very high levels of plasma HA or to other alterations, for instance, in the glycocalyx or the endothelium.
Our Hyal2–/– model joins a heterogeneous collection of lysosomal storage disorders termed the mucopolysaccharidoses that result from GAG accumulation. The first described hyaluronidase deficiency, resulting from HYAL1 mutations, was termed MPS IX (10
, 11
, 12)
. HYAL2 deficiency is now the second hyaluronidase deficiency to be described and may join the MPS IX group, Although this defect has yet to be described in humans, the relatively mild nature of the phenotype in the mice suggests that it would be compatible with life. With knowledge of the mouse phenotype for HYAL2 deficiency now available, it may be possible to identify this condition in humans as well. An accumulation of serum HA in the absence of the joint abnormalities identified in the HYAL1 patient could be a useful screening tool to identify potential patients for further testing for HYAL2 deficiency.
|
ACKNOWLEDGMENTS |
|---|
Received for publication April 23, 2008. Accepted for publication July 31, 2008.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. Duterme, J. Mertens-Strijthagen, M. Tammi, and B. Flamion Two Novel Functions of Hyaluronidase-2 (Hyal2) Are Formation of the Glycocalyx and Control of CD44-ERM Interactions J. Biol. Chem., November 27, 2009; 284(48): 33495 - 33508. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
