Correlation of Versican Expression, Accumulation, and Degradation during Embryonic Development by Quantitative Immunohistochemistry

Abstract

Versican, a chondroitin sulfate proteoglycan, is important in embryonic development, and disruption of the versican gene is embryonically lethal in the mouse. Although several studies show that versican is increased in various organs during development, a focused quantitative study on versican expression and distribution during lung and central nervous system development in the mouse has not previously been performed. We tracked changes in versican (Vcan) gene expression and in the accumulation and degradation of versican. Vcan expression and quantitative immunohistochemistry performed from embryonic day (E) 11.5 to E15.5 showed peak Vcan expression at E13.5 in the lungs and brain. Quantitative mRNA analysis and versican immunohistochemistry showed differences in the expression of the versican isoforms in the embryonic lung and head. The expression of Vcan mRNA and accumulation of versican in tissues was complementary. Immunohistochemistry demonstrated co-localization of versican accumulation and degradation, suggesting distinct roles of versican deposition and degradation in embryogenesis. Very little versican mRNA or protein was found in the lungs of 12- to 16-week-old mice but versican accumulation was significantly increased in mice with Pseudomonas aeruginosa lung infection. These data suggest that versican plays an important role in fundamental, overlapping cellular processes in lung development and infection.

Introduction

Versican, a component of the extracellular matrix and a member of the lectican family of proteoglycans, is the largest chondroitin sulfate proteoglycan and has an important role in the development of the heart, musculoskeletal system and central nervous system (CNS) (Milev et al. 1998; Mjaatvedt et al. 1998; Wight 2002; Shepard et al. 2007; Zhang et al. 2011; Hatano et al. 2012). The structure of versican is characterized by an approximately 550-kDa core protein consisting of an amino-terminal (G1) hyaluronan binding domain, a carboxy-terminal (G3) domain, and two central chondroitin sulfate glycosaminoglycan attachment domains, the α-GAG and β-GAG domains (Zimmermann and Ruoslahti 1989; Zako et al. 1995; Horii-Hayashi et al. 2008; Hatano et al. 2012). Versican has four different isoforms that differ in the central GAG attachment domain: V0 (contains both α and β GAG domains), V1 (contains only the β domain), V2 (contains only the α domain) and V3 (devoid of both GAG domains) (Zako et al. 1995; Gu et al. 2007).

Through interactions with hyaluronan and other extracellular matrix molecules and cell surface proteins, versican plays an important role in regulating cell functions such as cell adhesion, migration, proliferation, differentiation, apoptosis, morphogenesis and tissue hydration (Yamagata et al. 1993; Bensadoun et al. 1997; Wight 2002; Wu et al. 2005; Matsumoto et al. 2006; Sheng et al. 2006; Casini et al. 2008; Hatano et al. 2012). Versican isoform V1 promotes cell growth, aggregation and mesenchymal-epithelial transition, whereas the V2 isoform has an inhibitory effect in these processes (Schmalfeldt et al. 2000; Sheng et al. 2005; Sheng et al. 2006; Gu et al. 2007). The V2 isoform is highly expressed in the brain, predominately in white matter, with its expression increasing in the rodent brain after post-natal day (P)18 and remaining at the high level of expression seen in the adult brain (Milev et al. 1998; Oohashi et al. 2002).

Vcan expression during development is well characterized in the heart (Hatano et al. 2012) (Henderson and Copp 1998), and versican-deficient mice develop fatal cardiac defects by E10.5 (Mjaatvedt et al. 1998; Hatano et al. 2012). Changes in versican expression and distribution during embryonic development have also been reported in the CNS, where the V1 and V0 isoforms predominate during embryonic development and decrease rapidly at birth (Landolt et al. 1995; Milev et al. 1998; Schmalfeldt et al. 1998; Horii-Hayashi et al. 2008). Less is known about versican expression during lung development. Previous work in the mouse suggests peak versican gene expression at E13.5, with a subsequent decline over time to E18.5; the distribution of versican within the lung or the versican isoforms present at different stages of development were not reported (Rutter et al. 2010). A study in sheep showed versican distribution throughout the perialveolar region at all embryonic ages evaluated, with a decrease in versican lung expression in the perisaccular and alveolar regions during the last trimester of pregnancy associated with a reduction in lung tissue volume, suggesting an important role of versican in structural lung development (Faggian et al. 2007).

Versican and other proteoglycans are degraded by a family of A Disintegrin and Metalloproteinase with ThromboSpondin motifs (ADAMTS) proteinases (Porter et al. 2005; Kenagy et al. 2006; Stanton et al. 2011). Cleavage of V1 versican by ADAMTS-1, -4, -5 and -9 results in a 70-kD fragment containing the neoepitope peptide sequence, DPEAAE (Sandy et al. 2001; Russell et al. 2003; Porter et al. 2005; Longpre et al. 2009; Capehart 2010; Stanton et al. 2011). Versican expression has been shown to overlap and co-localize with ADAMTS and DPEAAE accumulation, suggesting increased proteolysis of versican during development (Kern et al. 2006; McCulloch et al. 2009). Little is known about DPEAAE accumulation in lungs and the CNS during development.

In a previous study, the highest level of Vcan expression in the mouse was detected at embryonic day (E) 13.5, and this level in the whole embryonic mouse was much higher than that observed in adult tissues (Naso et al. 1995). In adults, the highest expression of versican is found in the brain, with the lowest expression in the liver and intermediate expression in the lungs and heart (Naso et al. 1995). Versican expression, typically present at low levels in most adult tissues, is increased in the lung during experimentally induced inflammation (Gill et al. 2010; Chang et al. 2014). In human lungs, versican accumulation is increased in many pathological conditions (Bensadoun et al. 1996, 1997; Lowry et al. 2008; Merrilees et al. 2008; Gill et al. 2010; Chang et al. 2014; Wight et al. 2014a). Versican expression in the CNS is also increased following injury and altered in diseases such as multiple sclerosis (Fawcett and Asher 1999; Schmalfeldt et al. 2000; Sobel and Ahmed 2001; Asher et al. 2002; Gu et al. 2007).

The goal of this study was to determine Vcan expression and versican accumulation in the lungs and brain during embryonic development in the mouse. A 5-day time frame centered on E13.5 was chosen, as this day is a critical time period in rodent development and has previously been shown to have the highest Vcan expression both in the whole mouse embryo and in the developing mouse lung. Quantitative changes in versican accumulation and versican degradation products during this time period in the embryonic mouse lung and brain have not previously been reported. In addition, 12- to 16-week-old mice were infected with live P. aeruginosa for up to 5 days to measure changes in the accumulation of versican during lung infection. These studies show that 1) versican expression and accumulation peaks at E13.5 during development and 2) is then greatly decreased, with a minimal amount of versican found in the lungs of healthy 12- to 16-week-old mice, but is induced after exposure to live P. aeruginosa, with the highest accumulation of versican at 5 days post-infection.

Materials & Methods

Animal Model

C57Bl6/J wild type mice (The Jackson Laboratory; Bar Harbor, ME), housed under standard conditions and on a 12:12 hr dark:light cycle, were time-mated approximately 1 hr before the onset of the dark cycle and were checked for the presence of a vaginal plug the following morning. Embryos were considered to be gestational age 0.5 days (E0.5) on the morning that the vaginal plug was identified. E11.5–E15.5 (inclusive) embryos were harvested in the morning. Embryos were staged to ensure age-appropriate development, and embryos that did not appear developmentally age-appropriate were excluded (Kaufman 1992). Three litters for each time point were collected, with a median of 7 embryos per litter and 20 embryos per time point collected. From each litter harvested, three embryos were used for immunohistochemistry and the remaining embryos were used for mRNA isolation. Cerebrum, heart, liver, and lung samples were harvested from three healthy 8- to 12-week-old male and female C57Bl6/J mice and used as controls. All procedures were performed as part of an approved scientific protocol in accordance with the University of Washington Institutional Animal Care and Use Committee (IACUC).

Induction of Pseudomonas Aeruginosa Pneumonia

P. aeruginosa strain PAK, a nonmucoid, flagellated strain, obtained from Dr. Shawn Skerrett (Division of Pulmonary and Critical Care Medicine, Harborview Medical Center, University of Washington) through Dr. Stephen Lory (Department of Microbiology and Immunology, Harvard Medical School), was grown in 25 ml Trypic Soy broth at 37°C for 16 hr (stationary phase), collected, washed twice, and resuspended in 3 ml PBS. C57Bl6/N male mice (n=20) were infected with 10 million PAK in 50 µL PBS by oropharyngeal aspiration (Foster et al. 2001) under isoflurane anesthesia. Control mice (n=3) received PBS alone. Mice were sacrificed at 4 hr (n=6), 24 hr (n=5), 48 hr [n=6], and 5 days [n=3] post-infection. The left lung was snap-frozen in liquid nitrogen and the right lung was lavaged three times with PBS before inflation with 4% paraformaldehyde at 25 cm H₂O pressure, fixed overnight, washed with PBS twice, dehydrated in ethyl alcohol and embedded in paraffin. Five-micron tissue sections were cut for versican β-GAG immunohistochemistry.

Quantitative Real Time Reverse Transcription-PCR

Embryos used for mRNA isolation were dissected, and head, heart, lungs and liver from all of the embryos in the litter were pooled by organ for mRNA isolation, with the exception of heads from E13.5 and older embryos, which were processed separately due to their larger size. For E13.5 and later embryos, the face and snout was dissected from the head and removed before processing. Cerebral cortex, liver, heart and lung from three adult wild type C57Bl/6 mice were collected in RNAlater immediately following CO₂ euthanasia. Tissues for immunohistochemistry (brain, removed in its entirety and sectioned in the sagittal plane; liver and lungs), were also harvested from these mice.

RNA was extracted using RNAeasy Mini Kit as directed by the manufacturer (Qiagen; Valencia, CA) and cDNA was reverse transcribed using random primers with the High Capacity cDNA Reverse Transcription Kit, as directed by the manufacturer (Applied Biosystems; Foster City, CA). Quantitative real time reverse transcription polymerase chain reaction (PCR) was performed on an ABI Prism 7900HT Fast Real-Time PCR System using Taq Man Universal PCR Master Mix Reagents as directed by the manufacturer (Applied Biosystems). The ABI Gene Expression Assays used are as follows: total versican probe set, Mm00490179 or Mm01283063_m1;18S, Hs99999901_s1; V0 primer (MUV0, custom made), MUV0_F CCAAGTTCCACCCTGACATAAATGT, MUV0_RGGATGACCACTTACAATCATATCACTCA, MUV0_M FAM TCGACCTGTCTTGTTTTC; V1 primer, Mm00490173_m1; V2 primer (MUV2, custom made), MUV2_F CCAAGTTCCACCCTGACATAAATGT, MUV2_R GCATGGGTTTGTTTTGCAGAGATC, MUV2_M FAM CAGAGAAAACAAGACAGGACCT; V3 primer, Rn01493763_m1. The ABI Gene Expression Assay forward and reverse primers are proprietary. Normalized mRNA levels were then expressed as -fold of levels in adult control tissues using the comparative cycle threshold (Ct) method. The ΔCt was calculated as the difference in Ct values for the target genes as compared with 18S, and relative mRNA was calculated as 2^-ΔΔCt, as previously described (Livak and Schmittgen 2001; Chang et al. 2014).

Total versican PCR was performed twice per sample for the liver and heart and three times per sample for the lungs and head, with all samples on a single plate. Versican isoform PCR (V0 primer, V1 primer, V2 primer, V3 primer) was performed once for the lungs and head. For all samples, embryonic expression of versican was compared with the relative amount of versican in the average of the three adult samples.

In addition, versican copy number was determined by PCR and performed once with all of the samples on a single plate with the Versican V3 plasmid (mouse) total versican standard (generous gift of Kathleen Braun, Benaroya Research Institute) using serial 10-fold dilutions from 3×10⁵ copies to 30 copies, and a non-template control (0 copies/9 µl). A versican standard curve was constructed relating Ct value to versican copy number. Versican copy number using the V3 plasmid standard has previously been extrapolated to copy number using Ct values for V0, V1 and V2 (Shukla et al. 2010). For Fig. 1, copy number for the embryonic tissues based on the total versican probe set to measure mRNA Ct values for 15 ng cDNA was calculated according to this standard curve using the equation: Copy number=e^(-0.7482*Ct+29.4752).

Figure 1.

Cumulative Vcan expression by embryonic day. Contribution to the total embryonic gene expression by versican copy number in the lungs, liver, heart, and brain is shown at E11.5, E12.5, E13.5, E14.5, E15.5, and in the adult control. Changes in the amount of mRNA for versican was determined using mRNA collected from embryonic organ homogenates and quantitative real time PCR. A versican standard curve was constructed using the V3 plasmid (mouse) and probes for total versican related to Ct value for 15 ng cDNA to versican copy number. An asterisk (_*) shows groups that are significantly different (p<0.05) using a one-way ANOVA with Tukey’s multiple comparison test.

Versican and DPEAAE Quantitative Immunohistochemistry

Embryos for immunohistochemistry were removed from the embryonic membranes, initially fixed in 10% neutral-buffered formalin for 24–72 hr, and transferred to 70% ethanol before processing. Embryos were routinely paraffin embedded with three embryos per litter in one block in the same orientation and sectioned along the sagittal plane to include brain, heart, lungs and liver of the three embryos on the same slide. Lungs, liver and brain from 8- to 12-week-old adult mice harvested following euthanasia were routinely processed and paraffin embedded.

Versican immunohistochemistry was performed using an anti-α-GAG or an anti-β-GAG antibody, with reactivity to the V2, V0 and V1 versican isoforms based on the presence or absence of the GAG domains found for each isoform (V2 - α-GAG only, V0 - α-GAG and β-GAG, V1 - β-GAG only) (Milev et al. 1998; Popp et al. 2003; Snow et al. 2005; Horii-Hayashi et al. 2008; Smith et al. 2009). Immuno-histochemistry was performed as previously described (Fu et al. 2011; Chang et al. 2014). Briefly, all embryos were pretreated with Chondroitinase ABC for 1 hr and HIER1 (Citrate) for 10 min before combining with rabbit anti-mouse β-GAG antibody (#1033, lot #2006928; Millipore, Billerica, MA) for 1 hr at 2 µg/ml to detect for versican accumulation. Detection of α-GAG was performed on one litter (three individual mice) of E12.5 embryos and two litters (three mice per litter) of E13.5 embryos by immunohistochemistry using rabbit anti-mouse α-GAG antibody (#1032; Millipore) at 2 µg/ml for 1 hr following pretreatment with Chondroitinase ABC for 1 hr and HIER1 (Citrate) for 10 min (Supplemental Fig. 1). Detection of DPEAAE-reactive versican fragments was accomplished using a rabbit polyclonal antibody to versican αDPE Vc neoepitope (#PA1-1748A, lot#NE161982; Thermo Scientific, Wilmington, DE) at 2.5 µg/ml for 1 hr with HIER2 (EDTA) for 10 min. The primary antibodies used for immunohistochemistry were unconjugated. The primary antibodies were detected using the Bond automated stainer and Bond Refine Polymer Detection Kit (Leica Microsystems; Wetzlar, Germany). Negative controls were performed for studies of embryos of all gestational ages and 12 to 16- week old mice treated with PBS or live P. aeruginosa. Negative controls were performed using purified Rabbit IgG (#011-000-003; Jackson Immunoresearch Laboratories, West Grove, PA).

To provide quantitative measurements of versican accumulation and degradation, image analysis was performed following immunohistochemistry using whole-slide digital images and automated image analysis. Briefly, all slides were scanned in bright-field with a 20× objective using a Nanozoomer Digital Pathology slide scanner (Hamamatsu; Bridgewater, NJ). The whole-slide digital images were then imported into Visiopharm software (Hoersholm; Denmark) for analysis. For quantitative measurements of versican and DPAEE in embryos, the Visiopharm Image Analysis module was used to define regions of interest (ROIs) by manually drawing around visible area of lungs, liver, brain, and head. Brain and ventricle areas were measured, and the ventricle area was subtracted from the brain area. Head area was measured exclusive of the snout and soft tissue to approximate the sample obtained for mRNA analysis. The software converted the initial digital imaging into gray scale values using two features: RGB-R, with a mean filter of 5 pixels × 5 pixels, and an RGB-B feature. Visiopharm was then trained to label positive staining, versican β-GAG, or DPEAEE, and the background tissue counter stain (i.e., hematoxylin) using a project-specific configuration based on a threshold of pixel values. Images were processed in batch mode using this configuration to generate the desired outputs (area of versican, ratio of versican to total tissue area, area of DPEAAE, and ratio of DPEAAE to total tissue area). The ROIs were sampled at 100%. Any organ that was not visible on sagittal sections or showed substantial damage from processing was excluded from the analysis. For this reason, calculation of DPEAAE accumulation in lung was only available for one litter of embryos at E11.5 and two litters of embryos at E13.5. Based on output for each embryo analyzed, the mean value for each organ at each embryonic age was calculated by averaging the results for the individual embryos in each litter and then averaging the three litters from each embryonic age.

For 12- to 16-week-old mice that were treated with either PBS or live P. aeruginosa, glass slides of versican β-GAG immunohistochemistry were scanned in brightfield using the Hamamatsu NanoZoomer Digital Pathology System, as described for the embryos. The digital images were imported into Visiopharm software and ROIs were automatically drawn around lung tissue. The software converted the initial digital imaging into gray scale values using two features: RGB-G and Chromaticity Red feature. Visiopharm was then trained to label positive staining, versican β-GAG, and background tissue counter stain using a project-specific configuration based on a threshold of pixel values. Images were processed in batch mode using this configuration to generate the desired outputs. The ROIs were sampled at 100%. As for embryos, versican positively stained versus unstained tissues were segmented using a project-specific configuration to generate the desired outputs (area of lung tissue staining positive for versican; area of lung staining negative for versican; total area of lung tissue and ratio of versican positive lung tissue to total lung tissue area).

The Atlas of Mouse Development (Kauffman, 1992) and the Developing Mouse Brain Allen Brain Atlas (www.developingmouse.brain-map.org/static/atlas) were used to identify relevant anatomy for regions of positive versican and DPEAAE staining present on sagittal images.

Statistics

Statistics and images were generated in GraphPad Prism (La Jolla, CA). For PCR analyses of total versican gene expression, the fold increase over the average value of the adult samples was calculated for each embryo litter on two separate runs for liver and heart and three separate runs for head and lungs. Three litters per time point were used for E11.5, E12.5 and E15.5. Four litters per time point were used for E13.5 and E14.5. Values from each PCR run for fold increase over adult control of each embryo litter were averaged and the result was log transformed to yield an approximately normal distribution. For PCR analyses of the four versican isoforms, the fold increase over the average adult sample was calculated for each embryo litter on one run for lungs and head samples only. For β-GAG versican and DPEAAE quantitative immunohistochemistry analyses, the ratio of positive staining to total tissue area was calculated for each organ in each embryo, and values from mice in the same litter were averaged and log transformed to yield an approximately normal distribution. Age and organ comparisons for PCR and quantitative immunohistochemistry data were analyzed by one-way ANOVA followed by Tukey’s post-test with all possible comparisons performed. Statistical results with a p-value <0.05 were considered statistically significant.

Results

Expression of Versican mRNA in the Head, Heart, Liver and Lung from E11.5 to E15.5 compared with Adult Tissues

All of the embryonic tissues had higher levels of Vcan gene expression than the adult controls (Fig. 1 presents versican copy number by organ in embryo and adult tissues). In the adult, the liver had the least Vcan expression (27 copy numbers for 15 ng cDNA) and the brain had the most Vcan expression (1705 copy numbers for 15 ng cDNA). Vcan expression in the heart and lungs was intermediate, with 1435 copy numbers and 552 copy numbers per 15 ng cDNA, respectively. At E15.5, the pattern of Vcan expression for the organs tested was the same as that in the adult, with the highest expression in the brain, lowest expression in the liver, and intermediate expression in the heart and lungs (Fig. 1). The highest cumulative versican gene expression in the head, heart, liver and lungs occurred at E13.5 and was significantly higher at E13.5 than the cumulative versican expression at E14.5 or E15.5 (p<0.05; Fig. 1). Among the embryonic tissues, the liver had the least versican gene expression at every time point studied (Fig. 1). Vcan expression in the liver, as measured by copy number, was significantly lower than Vcan expression in the lungs, heart and head at E12.5, E13.5, E14.5 and E15.5 (p<0.05).

When expressed as a relative fold increase over the mean adult sample, Vcan expression in the liver varied significantly with time (p=0.0024) and Vcan expression decreased from E11.5 to E15.5 (p<0.05) (Supplemental Fig. 2). Vcan expression in the head was significantly higher at E13.5 than at E11.5 and E12.5 (p<0.05). Vcan expression in the lungs was also greatest at E13.5, although this difference was not statistically significant. However, when the lung analysis was restricted to the E11.5, E13.5, and E15.5 embryo samples, versican expression in the lung was significantly higher at E13.5 than at E15.5 (p<0.05).

Relative Gene Expression of Versican Isoforms V0, V1 and V2 in Embryonic Lungs and Head

Relative gene expression of the four versican isoforms compared with the healthy adult was evaluated in the lungs and the brain (Fig. 2). Relative mRNA expression (fold increase over adult) was highest for V0 and V1 at all time points both in the head and lungs. In the head, V0 and V1 isoform gene expression was higher at E13.5 than at E11.5 and E12.5 (p<0.05) and the V0 isoform expression in embryonic head was higher at E15.5 than at E11.5 (p<0.05). In the lungs, V0 (p=0.1380) and V1 (p=0.2959) isoform expression did not vary significantly over time; however, when the analysis was restricted to the E11.5, E13.5, and E15.5 embryo samples, both V0 and V1 versican expression was significantly higher at E13.5 than at E15.5 (p<0.05). Gene expression of the V2 isoform in lungs was similar to that of the adult control tissue and did not vary significantly at any embryonic age (Fig. 2D). However, the normalized Ct value for the V2 isoform in the embryonic heads varied significantly [adult Ct value = 30.7 for 15 ng cDNA (normalized Ct value = 19.5); E11.5 Ct value = 32.6 (normalized Ct value = 22.8) E13.5 and E15.5 Ct (and normalized Ct) values = 31.6 (20.1) and 31.1 (20.13), respectively, for 15 ng cDNA] and the relative V2 isoform expression in the head was higher at E13.5, E14.5 and E15.5 than at E11.5 (p<0.05) and higher at E13.5 than at E12.5 (p<0.05) (Fig. 2C). Relative gene expression of the V3 isoform did not vary significantly by embryonic age in the lungs but was significantly higher at E13.5 than at E12.5 in the head (p<0.05, data not shown).

Figure 2.

The V0 and V1 versican isoforms predominate in embryonic head and lungs from E11.5 to E15.5, and the V2 versican isoform increases in embryonic head but not lungs at E13.5. Changes in the relative amounts of mRNA for the four versican isoforms were determined using mRNA collected from embryonic organ homogenates and quantitative real time PCR. The expression of mRNA for the versican isoforms in the embryonic head (A) and lungs (B) is expressed as a relative fold increase in mRNA over the adult control. The relative expression of mRNA for the V2 isoform in the head (C) and lungs (D) is represented as a fold increase over the adult control. Values are the mean ± SEM with a minimum n=3 for each group studied. An asterisk (_*) shows groups that are significantly different (p<0.05) using one-way ANOVA with Tukey’s multiple comparison test. Ct values for V0 and V1 isoforms are indicated above E13.5 bars in head (A) and lung (B) and below E11.5, E13.5 and E15.5 bars for V2 in head (C) and lung (D).

Versican β-GAG and α-GAG Accumulation in the Embryonic Lung and Brain

β-GAG versican accumulation was seen throughout most of the organs at all embryo ages; although, consistently, there was very little immunoreactivity or Vcan mRNA in the liver. The liver therefore acted as an internal negative control for immunohistochemistry (Fig. 3A). At E11.5, there was strong staining of the mesenchyme of the lung bud (the scaffolding for lung development), with little positive staining of airway epithelium (Fig. 3B). At E13.5, very strong staining was present throughout the mesenchyme including the periphery of the lungs and the basement membrane, with little to no immunoreactivity of the airway epithelium (Fig. 3C). At E15.5, there was marked expansion of airways and alveolar space, with little immunoreactivity of the airway epithelial cells, strong positivity of the epithelial cell basement membrane, and moderate to strong immunoreactivity of the mesenchyme; although, areas of reduced immunoreactivity in the regions surrounding airways and along the outer surfaces of lungs were noted compared to earlier gestational ages (Fig. 3D). There was no immunoreactivity in tissues with the polyclonal negative control antibody (purified rabbit IgG) used for versican and DPEAAE immunohistochemistry (Supplemental Fig. 3).

Figure 3.

Versican β-GAG accumulation in embryonic lungs. Versican β-GAG accumulation in the whole embryo at E13.5 (A) and in the lung at E11.5 (B); E13.5 (C); and E15.5 (D). Brown indicates positive staining for versican β-GAG; blue, hematoxylin counterstain. AL, airway lumen; BV, blood vessel; C, choroid plexus; E, epithelium; H, heart; Lu, lung; Li, liver; M, medulla; NC, nasal cavity; S, spine; V, ventricle. Arrowhead indicates airway epithelial cell basement membrane. Scale (A) 1 mm; (B–D) 50 µm.

Pronounced β-GAG staining of the embryonic brain, mesenchyme of the developing skull and nasal process was present (Fig. 4A). In the brain at E11.5, there was a relatively large ventricle with very weak β-GAG immunoreactivity of the brain parenchyma surrounding the ventricle and slightly stronger β-GAG immunoreactivity of the deeper mantle zone (Fig. 4B). At E13.5, the choroid plexus was apparent and little immunoreactivity in the cells of the choroid plexus was seen (Fig. 4C). There was an increasingly laminar appearance to the cortex at E13.5 with immunoreactivity of the roof (superficial stratum) of the neopallial cortex, roof of the midbrain, and the developing olfactory lobe (Fig. 4A). There was very strong, homogeneous staining of the upper lip and palatal shelf of the maxilla (Fig. 4A). At E15.5, increasing convolution and organization of the neural tube and ventricular system was apparent (Fig. 4D). There was strong staining of the cerebellar primordium at E15.5 and portions of the periventricular medulla and midbrain (Fig. 4D). In the adult brain, there was little positive β-GAG versican staining, with the most staining present in the granular layer of the cerebellum and mild staining in the cerebellar white matter (Supplemental Fig. 1).

Figure 4.

Versican β-GAG accumulation in embryonic brain. Versican β-GAG accumulation in the embryonic head at E13.5 (A) and in the brain at E11.5 (B); E13.5 (C); and E15.5 (D). Brown indicates positive staining for versican β-GAG; blue, hematoxylin counterstain. C, choroid plexus; CS, corpus striatum; D, diencephalon; FV, fourth ventricle; M, medulla; MB, midbrain; NC, nasal cavity; NP, nasal process; O, olfactory lobe; P, pons; R-HB, roof of hindbrain; R-MB, roof of midbrain; R-NC, roof of neopallial cortex; TV, telencephalic vesicle; UL, upper lip; V, lateral ventricle. Scale (A, D) 1 mm; (B) 500 µm; (C) 100 µm.

Quantitative immunohistochemistry, like Vcan mRNA expression, was highest in the lung at E13.5 (Fig. 5A, 5B). There was greater β-GAG versican immunoreactivity of the lungs at all embryo ages as compared with that in the adult control (Fig. 5B; p<0.05). There also was greater β-GAG versican immunoreactivity of the brain at E13.5, E14.5, and E15.5 as compared with that in the adult control (p<0.05), which was similar to Vcan expression in the head (Fig. 5C, 5D). The embryonic brain and head showed similar patterns of β-GAG versican immunoreactivity over time, with the head consistently higher than the brain (data not shown) and the highest value at E13.5. Very little positive β-GAG versican accumulation was noted in the adult lung (0.05% of total tissue area), brain (1.8% versican positive staining to total tissue area), and liver (0.4% of total tissue area).

Figure 5.

Vcan expression and β-GAG accumulation in lungs and brain from E11.5 to E15.5. Versican relative gene expression (A, C) and β-GAG quantitative immunohistochemistry (B, D) in embryonic lungs (A, B) and head (C, D) over five embryonic ages. Changes in versican immunohistochemistry were determined by calculating ratios of versican to total tissue area using the Visiopharm image analysis module. Changes in relative amounts of mRNA for versican were determined using mRNA collected from embryonic organ homogenates and quantitative real time PCR using total versican primer sets. Values are the mean ± SEM. mRNA for versican is expressed as a relative fold increase in mRNA over the adult control. An asterisk (_*) shows groups that are significantly different (p≤0.05).

In the whole embryo at E12.5 and E13.5, there was little α-GAG versican accumulation in the lung, with only mild immunoreactivity of the cells lining the airways (Supplemental Fig. 1). In the embryonic brain at E13.5, there was strong α-GAG as well as β-GAG versican immunoreactivity throughout the roof of the neopallial cortex, roof of the midbrain, and deeper portions of the hindbrain, with less immunoreactivity in the tissue surrounding the ventricle (Supplemental Fig. 1). Portions of the medulla and pons showed stronger α-GAG than β-GAG immunoreactivity. Portions of the spinal column were strongly positive on α-GAG but not β-GAG immunohistochemistry (Supplemental Fig. 1). The choroid plexus and portions of the nasal cavity and maxilla showed stronger β-GAG than α-GAG immunoreactivity (Supplemental Fig. 1).

Relationship between the Accumulation of the Versican Degradation Product, DPEAAE, and Versican β-GAG Accumulation as Measured with Immunohistochemistry

At E13.5, there was mild DPEAAE immunoreactivity of the lung mesenchyme, which co-localized with regions of decreased versican accumulation (Fig. 6A, 6B). At E15.5, DPEAAE immunoreactivity appeared somewhat stronger around the basement membrane and the external surface of the pulmonary parenchyma, which co-localized with areas of decreased versican accumulation at these ages (Fig. 6C, 6D). Image analysis showed no significant difference in DPEAAE immunoreactivity over time in the lungs (p=0.6788; Fig. 6E), although it appeared that, at embryonic ages when the versican immunoreactivity was greatest, the DPEAAE immunoreactivity was lower. Correlation between versican β-GAG accumulation and DPEAAE accumulation for the embryonic lung is presented (Fig. 6F).

Figure 6.

Areas of staining in the embryonic lungs for versican and DPEAAE immunohistochemistry (IHC) colocalize. (A) E13.5 versican β-GAG IHC (40×). (B) E13.5 DPEAAE IHC (40×). (C) E15.5 versican β-GAG IHC (20×). (D) E15.5 DPEAAE IHC (20×). (E) DPEAAE accumulation in the lung from E12.5 through E15.5 by quantitative IHC. (F) Correlation between β-GAG versican and DPEAAE accumulation on quantitative IHC. Brown indicates positive staining for versican β-GAG or DPEAAE; blue, hematoxylin counterstain. Arrows indicate regions of decreased versican immunoreactivity (A, C) and increased DPEAAE immunoreactivity (B, D). Asterisks (*) indicate regions of decreased versican immunoreactivity. Arrowhead indicates basement membrane. AL, airway lumen; E, airway epithelium. Scale (A, B) 50 µm; (C, D) 100 µm.

In the developing brain, areas of positive staining on DPEAAE IHC co-localized with regions of positive staining on β-GAG versican IHC in corpus striatum (Fig. 7A, 7B) and the medulla at E15.5 (Fig. 7C, 7D). There was mild, diffuse positive DPEAAE staining at E11.5 and E12.5 (data not shown). At E13.5, there was mild, positive immunoreactivity in the roof of the neopallial cortex and portions of the corpus striatum, midbrain, cerebellar primordium, and medulla. At E15.5, mild staining of the roof of the neopallial cortex and midbrain was also noted (data not shown). Much weaker DPEAAE staining than versican β-GAG staining was seen in the developing cartilage, nasal cavity and jaw (data not shown). In the brain, DPEAAE accumulation decreased from E11.5 to E13.5 (p<0.05; Fig. 7E). As for the lung, when the versican immunoreactivity was greatest at E13.5, the DPEAAE immunoreactivity was lower. Correlation between versican β-GAG accumulation and DPEAAE accumulation for the embryonic brain is presented (Fig. 7F).

Figure 7.

Areas of staining in the embryonic brain for versican and DPEAAE immunohistochemistry (IHC) colocalize. (A) E15.5 versican β-GAG IHC. (B) E15.5 DPEAAE IHC. (C) E15.5 versican β-GAG IHC. (D) E15.5 DPEAAE IHC. (E) DPEAAE accumulation in brain from E11.5 through E15.5 by quantitative IHC. (F) Correlation between β-GAG versican and DPEAAE accumulation on quantitative IHC. Brown indicates positive staining for versican β-GAG (A, C) and DPEAAE (B, D); blue, hematoxylin counterstain. Black arrows indicate regions of both increased DPEAAE and β-GAG versican immunoreactivty. Asterisks (*) indicate regions of increased β-GAG versican but decreased DPEAAE immunoreactivity. C, choroid plexus; CP, cerebellar primordium; CS, corpus striatum; D, diencephalon; M, medulla; P, pons; V, ventricle. Scale (A, B) 100 µm; (C, D) 500 µm.

Virtually no DPEAAE accumulation was seen in adult brain, heart, lungs, or liver (all tissues <0.1% ratio of DPEAAE-positive expression to total tissue area).

Versican Expression Is Increased in the Adult Lung in Pseudomonas Aeruginosa Pneumonia

The healthy lung is a highly quiescent tissue. In contrast, injured lungs have remarkable reparative capacity, which raises the question as to whether similar mechanisms regulate cellular differentiation and proliferation in lung development and repair (Beers and Morrisey 2011; Kotton and Morrisey 2014). Our previous work showed increased versican expression in mice exposed to gram-negative bacteria; however, these studies did not measure versican accumulation (Chang et al. 2014). To measure changes in versican accumulation following infection, 12- to 16-week-old mice were exposed to the gram-negative bacteria, P. aeruginosa. Following oropharyngeal aspiration of P. aeruginosa, there was a significant increase in the accumulation of β-GAG-positive versican staining in the total lung by quantitative immunohistochemistry at 48 hr and 5 days (p<0.05) as compared with lungs from PBS-treated control mice (Fig. 8). Positive staining occurred within macrophages in the pulmonary interstitium and the alveoli (Fig. 8C, inset), in the alveolar septa, and in the perivascular space around airways and blood vessels. This area also stained positively for carbohydrate with Periodic acid-Schiff staining (Supplemental Fig. 4). Positive immunostaining in the lungs of mice infected with P. aeruginosa was consistently localized to regions where inflammatory cells (i.e., macrophages, lymphocytes and neutrophils) were located (Fig. 8). In contrast, no increase in α-GAG staining was noted in P. aeruginosa-treated mice over control mice on quantitative immunohistochemistry; although, some nonspecific α-GAG immunoreactivity in the control and infected lungs was noted (data not shown).

Figure 8.

Versican accumulation during embryonic development in healthy adult lungs and in lungs of a mouse with Pseudomonas aeruginosa lung infection. (A) Versican accumulation in fetal lung tissue at E14.5 days. (B) Versican accumulation in the lung of a 16-week-old mouse treated with PBS as a vehicle control. (C) Versican accumulation from a 16-week-old mouse infected with live Pseudomonas aeruginosa for 5 days. Brown indicates positive staining for versican β-GAG; blue, hematoxylin counterstain. Br, bronchiole; Di, diaphragm; Ri, rib; PV, postcapillary Vein; TB, terminal bronchiole. Arrows indicate Vcan staining in the alveolar septa; * marks an area of positive staining of the alveolar septa; cells in alveolar space makes it difficult to distinguish these two anatomical compartments. (D) The amount of versican-stained lung tissue as a percentage of total lung tissue in control mice (PBS) and those exposed to live P aeruginosa for up to 5 days. Values are the mean ± SEM (n=3 to 6). ^asignificantly different from PBS, ^bsignificantly different from 4 hr, ^csignificantly different from 24 hr. p<00001 using a one-way ANOVA with Tukey’s multiple comparison test. Scale (A–C) 100 µm; (C inset) 50 µm.

Discussion

The goal of this study was to determine Vcan expression and versican accumulation in the lungs and brain during embryonic development in the mouse. We found that E13.5 is the time of highest total versican expression and appears to be a critical time for versican expression in both rodent lung and brain development. This increased expression of versican occurs at a time of increased tissue complexity and differentiation of these organs. In addition to measuring versican expression, we performed immunohistochemistry and image analysis to provide objective spatial and temporal information about versican accumulation in the lungs and brain. To our knowledge, quantitative measurements of versican accumulation and versican degradation products during this time period in the embryonic mouse lung and brain have not been reported. We found that areas of versican accumulation in the lungs and brain generally involved regions of cell proliferation and differentiation and locations of scaffolding for tissue development and migration, such as the pulmonary interstitium and the developing roof of the neopallial cortex and hindbrain. These areas of versican accumulation co-localized with the accumulation of the versican degradation product DPEAAE. Similar to that seen in the adult, at the latest embryonic time point tested (E15.5), versican gene expression was lowest in the liver, highest in the brain, and showed intermediate expression in the heart and lungs. Based on our finding of low versican expression in the healthy, normal adult mouse lung, we wanted to determine if versican was an important developmental molecule that was reactivated upon injury. Mice exposed to P. aeruginosa for 48 hr or 5 days had a significant increase in the accumulation of versican in the lungs, which suggested a critical role for this proteoglycan in lung inflammation as well as in the resolution of lung inflammation and injury.

We showed maximum Vcan expression in embryonic lungs at E13.5, a time of pulmonary vascularization, right lung lobation, differentiation and expansion of mesenchymal progenitor cells, and airway epithelialization, with a subsequent decline at later embryonic ages, consistent with results of previous studies in mice and sheep (Faggian et al. 2007; Rutter et al. 2010; Hung et al. 2013). We saw strong versican β-GAG accumulation at all ages examined throughout the mesenchyme of the developing lungs, with pronounced positive staining of the basement membrane of airways but little to no accumulation in airway epithelium. The immunoreactivity throughout the lung mesenchyme suggests a key role of versican in differentiation and expansion of mesenchymal progenitor cells into pericytes, vascular smooth muscle cells and endothelial cells (Hung et al. 2013; Sims-Lucas et al. 2013). At E14.5 and E15.5—the time when a pronounced increase in bronchioles per region of lungs occurs—there was reduced β-GAG accumulation around the airways and along the periphery of the lungs. This decreased β-GAG immunoreactivity co-localized with areas of increased DPEAAE immunoreactivity. Colocalization of DPEAAE and versican immunoreactivity during embryonic development of the mouse limb and heart has been previously reported (Kern et al. 2006; Kern et al. 2007; Capehart 2010). Colocalization of ADAMTS5 and versican immunoreactivity during embryonic development in the mouse lung at later stages of embryonic development (E16.5) has also been previously reported (McCullough et al. 2009). With quantitative immunohistochemistry at E13.5, there was little DPEAAE accumulation in the lungs, suggesting antagonistic roles of versican deposition and versican proteolysis in promoting and inhibiting cellular growth and migration through a provisional extracellular matrix during development, as has been previously proposed in lung, cardiac and musculoskeletal system development and in lung inflammation (Kern et al. 2006; Capehart 2010; Nandadasa et al. 2014).

Most previous work involving versican expression in the developing rodent CNS focuses on later gestation (>E16); however, the results of our study suggest an important role of versican at earlier times. In our study, versican gene expression in the embryonic head was significantly higher at E13.5, the time of neopallial cortex and olfactory lobe expansion and differentiation, as compared with that seen at earlier embryonic ages. With immunohistochemistry, versican β-GAG accumulation was most pronounced in the roof of the neopallial cortex and midbrain and portions of the pons and medulla, and the highest versican β-GAG accumulation occurred at E13.5, consistent with our mRNA results. At later ages (E13.5, E14.5 and E15.5), we noted β-GAG versican accumulation in the roof of the midbrain, cerebellar primordium, and developing olfactory lobe, consistent with previous findings in the rat at E16 (Popp et al. 2003). Also at later ages (E14.5 and E15.5), we found increased DPEAAE accumulation, which co-localized with regions of β-GAG versican accumulation in the developing cortex and medulla. In other regions of the embryonic brain, such as the periventricular stratum of the medulla and midbrain, positive β-GAG versican staining occurred in areas of negative DPEAAE staining. This may reflect dynamic versican degradation, which occurs to limit further migration and cellular differentiation once the process is complete, as previous work suggests that not only is there ADAMTS-mediated clearance of versican at specific embryonic times, but also a DPEAAE degradation product of versican that may promote local apoptosis and cell proliferation (Nandadasa et al. 2014).

One limitation of the current study is that versican gene expression in the embryonic head was used as an estimate of gene expression in the brain. Therefore, when quantitative PCR was performed, some developing bone and skin was analyzed in addition to the nervous tissue. Previous studies show high Vcan expression in human fetal skin, which may have affected our results (Carrino et al. 2011). However, on our quantitative immunohistochemistry analysis, we measured head and brain separately and found a similar pattern for all of the embryonic ages studied, although versican immunoreactivity as a percentage of total tissue was consistently less for the brain than for the whole head. This similarity between Vcan gene expression in embryonic head and versican accumulation in brain during development suggests that the embryonic head serves as a reasonable approximation of embryonic brain for our mRNA data. Another limitation is that the entire embryonic organ was used for gene expression assays and organs from littermates were pooled, while a representative portion of the organ was used from the adult. This may have led to discrepancies in versican content depending on the location sampled in an organ, especially for less homogenous organs such as the brain. However, we were able to assess versican immunoreactivity throughout a sagittal section of the adult brain. The use of sagittal sections limited our ability to identify specific anatomical structures and tracts within the brain; however, the sagittal section provided an acceptable overview of versican accumulation in the brain and lung in a single slice for quantitative immunohistochemistry.

Our studies in 12- to 16-week-old mice exposed to PBS or live P. aeruginosa showed a significant increase in the accumulation of versican in infected lungs, which localized to the alveolar septa and regions with increased numbers of leukocytes. A number of in vitro studies suggest that versican plays a key role in the immune response to lung infection by modulating leukocyte migration, adhesion, and activation (Hirose et al. 2001; Wu et al. 2005; Li et al. 2008; Potter-Perigo et al. 2010; Evanko et al. 2012; Said et al. 2012; Wight et al. 2014a). In vitro studies also suggest that versican plays a role in the resolution of lung inflammation and injury through regulation of cellular differentiation and proliferation (Evanko et al. 1999; Sheng et al. 2005; Wight et al. 2014b). Therefore, our findings of high versican expression at 5 days post-infection, coinciding with recovery from injury, together with in vitro studies, suggests that versican plays a role in modulating key processes that are required to return the inflamed and injured lung to a healthy state.

Based on the observation that versican is expressed in many tissues during embryogenesis, it is not surprising that the disruption of the versican gene is embryonically lethal (Mjaatvedt et al. 1998). Unfortunately, the lack of readily available versican-null mice means that the majority of our knowledge about versican biology comes from in vitro studies. The observations that the expression of versican is maximal at E13.5, is present at very low amounts in tissues of healthy adults, and is reactivated in lungs in response to inflammation, suggests that versican may play an important role in many fundamental cellular activities that occur during development and that overlap with processes that occur in acute and chronic inflammation (Bensadoun et al. 1996, 1997; Asher et al. 2002; Wight 2002; Kenagy et al. 2006; Merrilees et al. 2008; Beers and Morrisey 2011; Zhang et al. 2011; Warburton 2012).

In conclusion, we present data showing excellent agreement between patterns of Vcan gene expression and versican accumulation during embryonic development of lungs and brain. Vcan expression was highest at E13.5 in the lungs and the brain, an embryonic age in the mouse associated with rapid expansion and differentiation of these two organs. Versican degradation, as measured by DPEAAE immunoreactivity, tended to be low when versican β-GAG immunoreactivity was high in these organs at E13.5 and increased at the later gestational ages, following the peak Vcan expression at E13.5. In addition, we found that, in healthy adult lungs, there is little versican mRNA or accumulation, but versican accumulation is increased in the lungs of mice following exposure to live P. aeruginosa. These findings suggest that versican plays an important role in fundamental, overlapping cellular processes in lung development and infection. Future work using mice harboring a disrupted versican gene via conditional knockout strategies will be critical to identify the role versican plays in embryogenesis, health and disease.