CD34 protein is expressed in murine, canine, and porcine lungs

Abstract

The cell surface protein CD34 is expressed in various human tissues and cells, including hematopoietic stem cells, vascular endothelial cells, mucosal dendritic cells, mast cells, eosinophils, microglia, fibrocytes, muscle satellite cells, and platelets. There is a lack of data on the expression of CD34 in canine and porcine tissues. Therefore, we designed a series of immunoblotting, immunohistochemistry, and immunofluorescence experiments to observe CD34 expression in murine, canine, and porcine lungs. We used a rabbit antibody (clone EP373Y) to target the conserved human CD34 C-terminal region and validated its immunoreactivity against mouse lung homogenates. The data showed diffuse bronchiolar and alveolar epithelial localization of CD34 protein in normal murine, canine, and porcine lungs. At 9 or 24 h after bacterial endotoxin exposure, murine CD34 protein shifted to specific bronchoalveolar cells with a punctate pattern, as quantified by CD34 fluorescence. Specific porcine bronchoalveolar cells and leukocytes had significant CD34-positive immunostaining after H3N1 influenza infection. Thus, our study provides fundamental data on the expression of CD34 in lungs and validates an antibody for use in further experiments in these animal species.

Introduction

CD34-family proteins are proposed to play various roles depending on the site of expression (1). The cell surface protein CD34 has been reported in multiple human tissues and cells. The study of CD34 biology has been facilitated through the use of 2-color fluorescence-activated cell sorting to isolate CD33⁻CD34⁺ cell populations from hematopoietic stem cells (2). Long-term marrow cultures of CD33⁻CD34⁺ cells have generated colony-forming cells for more than 5 wk. Vascular endothelial cells (3,4), mucosal dendritic cells (5), mast cells (6), eosinophils (7,8), microglia (9), fibrocytes (10), muscle satellite cells (11), and inactivated platelets (12) have expressed CD34. Fibroblast-like cells situated in the submucosa flanked by smooth muscle cells of the muscularis mucosae and the muscularis propria have stained positive for CD34. CD34 has also been expressed in the lamina propria’s endothelial cells in the cecum of C57Bl/6 mice (13). In dogs, CD34 expression has been observed in hematopoietic stem cells in bone marrow (14), peripheral blood (15), and hair follicles (16).

The cytoplasmic region of CD34, encoded by exons 7 and 8, contains the highest degree of sequence similarity between human, murine, and canine CD34. There is greater than 90% amino acid identity and approximately 92% nucleotide identity within this region. This region contains several known or potential protein kinase phosphorylation sites. If we compare the entire human and canine CD34 proteins, there is 70% amino acid identity between the two.

We validated and used a rabbit monoclonal antibody (mAb) (clone EP373Y) to assess the localization of CD34 in normal murine, canine, and porcine lungs. In addition, we examined the inflamed lungs of mice and pigs for the localization of CD34.

Materials and methods

Animals

Lung samples were collected from normal mice (n = 3) and those treated intranasally with Escherichia coli lipopolysaccharide (50 μg in 50 μL saline) (n = 3). Paraffin-embedded normal (n = 4) and H3N1 influenza-infected (n = 4) pig lung tissue blocks were kindly provided by Dr. John Harding from the Western College of Veterinary Medicine, University of Saskatchewan. Healthy canine lung tissues (n = 4) were obtained from the Society for the Prevention of Cruelty to Animals in Saskatoon, Saskatchewan.

Western blot

Western blots were performed on wild-type (WT) and CD34 knock-out (KO) mice and dog lung homogenates. Protein extracts were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene difluoride membranes. These were probed with antibodies against CD34 (rabbit mAb, cat no. ab81289, 1:10 000; Abcam, Cambridge, Massachusetts, USA) and β-actin (mouse mAb, cat no. ab8224, 1:1000 for dog homogenates and 1:5000 for mouse homogenates; Abcam). The mouse CD34 proteins were detected by chemiluminescence detection (GE Healthcare, Chicago, Illinois, USA) on a ChemiDoc MP Imagelab (Bio-Rad Laboratories, Hercules, California, USA) of goat anti-rabbit immunoglobulin G (IgG), horseradish peroxidase (HRP) conjugate. Membranes were re-probed for β-actin after incubating the membranes with proprietary blot re-store solution (Millipore, Temecula, California, USA) and goat anti-rabbit IgG, HRP conjugate. The canine proteins of interest were detected with fluorescein-conjugated bovine anti-mouse IgG-fluorescein isothiocyanate antibody (sc-2366, 1:1000; Santa Cruz Biotechnology, Dallas, Texas, USA) and Cy5-conjugated goat anti-rabbit IgG antibody (ab97077, 1:1000; Abcam).

Immunohistochemistry

Tissue sections were deparaffinized with xylene and rehydrated in an ethanol series. Endogenous peroxidase activity was inactivated with 0.5% H₂O₂ in methanol in the dark for 20 min at room temperature. After washing, antigen retrieval was performed by incubating sections with warmed pepsin (2 mg/mL in 0.01 N HCl) at 37°C for 40 min. After washing, they were blocked with 1% bovine serum albumin in 1% phosphate-buffered saline (PBS) for 30 min at room temperature. The sections were then incubated overnight at 4°C with primary antibodies against CD34 (rabbit mAb EP373Y, ab81289, 1:500; Abcam) and von Willebrand factor (vWF) (A0082, 1:300; Dako Denmark A/S, Glostrup, Denmark). After washing, the tissue sections were incubated with secondary antibody (HRP-conjugated goat anti-rabbit IgG antibody, P0448, 1:500, Dako Denmark A/S, Glostrup, Denmark) for 30 min. The sections were developed with VECTOR VIP Peroxidase (HRP) Substrate Kit (SK-4600; Vector Laboratories, Burlingame, California, USA) for 5 min, counterstained with methyl green, and mounted with cover slips.

Immunofluorescent staining

Tissue sections were deparaffinized with xylene and rehydrated in an ethanol series. Endogenous peroxidase activity was inactivated with 0.5% H₂O₂ in methanol in the dark for 20 min at room temperature. After washing, antigen retrieval was performed in 2 steps. First, heat-induced epitope retrieval was performed by incubating sections in sodium citrate buffer (10 mM, pH 6.0) at 90°C to 95°C for 20 min. Second, the sections were allowed to cool in distilled water, following which they were incubated with warmed pepsin (2 mg/mL in 0.01 N HCl) at 37°C for 20 min. After washing, they were blocked with 1% bovine serum albumin in 1% PBS for 30 min at room temperature. The sections were then incubated overnight at 4°C with 100 μL primary antibodies per section against CD34 (rabbit anti-mouse, EP373Y, ab81289, 1:500; Abcam). The next day, the slides were left at room temperature for 30 min followed by washing with PBS × 3 times. After washing and incubation with 100 μL secondary antibody per section (Alexa Fluor 488 conjugated anti-rabbit IgG, green; (ThermoFisher, Carlsbad, California, USA) for 30 min in the dark, sections were washed and counterstained with 4′,6-diamidino-2-phenylindole (DAPI). Sections were allowed to dry and cover slips were mounted with ProLong Gold mounting medium (ThermoFisher). Media was cured for 24 h before sealing the cover slip with nail varnish. Stained slides were stored in the dark at 4°C.

Image analysis

Lung immunofluorescence images (16-bit, 2-color) were processed for DAPI and CD34 protein quantification using Fiji (https://fiji.sc/), an open platform image processing software. The bronchiolar, vascular, and alveolar septal regions from the DAPI-stained images were outlined in yellow and numbered. The regions were copied onto the corresponding CD34 stained image. Image parameters such as area, perimeter, sum of fluorescence intensity (FI), skewness, and kurtosis were quantified for each channel (DAPI and CD34). We computed 4 image parameters and calculated the sum of CD34 FI to sum of DAPI FI ratio to use as a surrogate for the total protein expression per cell. The ratio of the sum of CD34 FI per 100 μm of lung tissue was a surrogate for the area normalized protein expression. CD34 skewness indicated the frequency distribution of CD34 in regions analyzed and CD34 kurtosis provided information about the spatial distribution of CD34 protein in the analyzed regions. This procedure was repeated for similarly stained lung sections from 3 separate animals across groups (i.e., 0, 9, and 24 h post LPS exposed mouse lungs).

Statistical analysis

Murine lung CD34 image expression parameters are graphically represented for each group in scatterplots. For the analysis, data from 3 separate animals (i.e., n = 3) are expressed as mean ± standard error of the mean (SEM).

Data were tested for normality/lognormality using the Shapiro-Wilk test, before conducting a 1-way analysis of variance (ANOVA). The data for CD34 kurtosis was neither normal nor lognormal. Thus, a Kruskal-Wallis test was used for overall analysis of variance followed by Dunn’s pairwise comparisons. Statistical significance was set at P < 0.05. All data were analyzed using the GraphPad Prism version 8 software package (San Diego, California, USA).

Results

CD34 localizes in the murine bronchiolar and alveolar epithelium as well as the vascular endothelium

Sodium dodecyl sulphate–polyacrylamide gel electrophoresis revealed that the antibody raised against the CD34 C-terminal was able to specifically detect a single 102 to 150 kDa CD34 protein band in WT mice (Figure 1). The antibody showed no immunoreactivity in the lung homogenates from KO mice.

Figure 1.

Validation of CD34 antibody. The CD34 antibody was validated against wild-type (WT) and knock-out (KO) mice lung homogenates by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Representative western blots from WT and KO mouse lung homogenates, blotted against the rabbit anti-CD34 antibody (ab81289 clone EP373Y; 1:10 000), and after stripping, blotted against rabbit anti β-actin antibody (ab8227; 1:5000) (n = 3). The images shown are cropped and adjusted for brightness and contrast based on antibody binding and molecular mass standards (102 to 150 kDa CD34 and 40 to 50 kDa β-actin band). Note that the antibody does not cross-react at all in homogenates of KO mice.

In normal murine lungs, CD34 was expressed in the bronchial epithelial cells, the vascular endothelium, and the alveolar septal cells (Figures 2A to 2C, C′, and Supplementary Movie 1; available from the authors upon request). The airway epithelial cells showed intense CD34 reaction on their apical surfaces. Staining in the cytoplasm and the basolateral surfaces was also noticed. At 9 (Figures 2D to 2F, F′) and 24 h (Figures 2G to 2I, I′) after LPS exposure, we observed punctate staining in alveolar septa and the endothelium of small pulmonary vessels. The staining aggregates in the septa were likely the recruited inflammatory cells expressing CD34 (Figure 2F, F′). The CD34 staining in the airway epitheliums of LPS-exposed mice also changed to a more concentrated pattern in bronchiolar epithelial cells (Figures 2G to 2I, I′) compared to the diffuse cellular staining in airways of normal mice (Figure 2C, C′).

Figure 2.

Murine lung CD34 immunofluorescence staining. Representative mouse lung immunofluorescence cryosections stained with DAPI shown in blue (A, D, G) and rabbit anti-CD34 antibody shown in green (ab81289 clone EP373Y) (B, E, H). C, F, and I show merged panels for murine lungs stained at baseline (0 h), 9 h post-LPS exposure, and 24 h post-LPS exposure, with respective insets (C′, F′ and I′) showing higher magnifications of CD34 positive cells. Note the strong apical and border distribution of CD34 across bronchiolar epithelial and septal cells at baseline (C′). Lungs from LPS treated mice show a punctate staining pattern at 9 h (F′), which becomes diffuse at 24 h (I′).

B — bronchus; BV — blood vessel.

Next, we quantified the immunofluorescent expression of CD34 in lung tissues of WT mice without (0 h) or 9 and 24 h after LPS exposure (Figures 2A to 2I). As described in the methods and shown in Supplementary Figure 1A–I, lung regions were outlined to mark bronchiolar, alveolar, or vascular compartments and CD34 FI was quantified with respect to DAPI FI (Figure 3A). We did not observe a change in CD34 FI with respect to DAPI expression, which is likely due to the compensatory increase in alveolar CD34 compared to bronchiolar FI after LPS exposure. However, CD34 FI per 100 μm of lung parenchyma was significantly more at 0 h compared to 9 or 24 h post-LPS exposure (P < 0.01) (Figure 3B). A detailed analysis of the pixel frequency distribution revealed that CD34 pixels frequently skewed towards higher intensities at both 9 (P < 0.05) and 24 h (P < 0.05) after LPS exposure, when compared to 0 h (Figure 3C). Moreover, the kurtosis of the CD34 staining pattern was discrete or heterogeneous at 9 h (P < 0.06) post-LPS exposure when compared to the homogenous staining pattern at 0 h (Figure 3D).

Supplementary Figure 1.

Murine lung CD34 immune-fluorescent staining. Representative mouse lung immune-fluorescent cryo-sections outlined for regions of interest for calculation of image parameters shown in figure 3 for (A, D, G) DAPI shown in blue and (B, E, H) rabbit anti-CD34 antibody shown in green (ab81289 clone EP373Y). (C, F, I) show merged panels for murine lungs stained at baseline i.e., 0 h, 9 h post-LPS exposure and 24 h post-LPS exposure. Note the strong apical and border distribution of CD34 across bronchiolar epithelial and septal cells at baseline (C′). Lungs from LPS treated mice show punctate staining pattern at 9 h (F′) which becomes diffuse at 24 h (I′).

B — Bronchus; BV — Blood vessel.

Figure 3.

Immunofluorescence analysis was performed on regional lung outlines overlaid in 4′,6-diamidino-2-phenylindole (DAPI) and CD34 channels to quantify lung CD34 image parameters. A — Ratio of CD34 fluorescence intensity (FI) to DAPI FI, quantified at 0, 9, and 24 h post-LPS exposure as an index of the brightness of CD34 for each cell. B — CD34 FI per 100 μm of lung tissue, quantified at 0, 9, and 24 h post-LPS exposure as an index of the CD34 brightness within 100 μm across groups. C — Frequency distribution of lung CD34 at 0, 9, and 24 h post-LPS exposure measured by the skewness of CD34 positive pixels (e.g., greater degree of skewness means brighter or higher intensity pixels). D — Spatial distribution of lung CD34 at 0, 9, and 24 h post-LPS exposure measured by the kurtosis of CD34 positive pixels (e.g., higher values denote discrete or punctate distribution). Murine lung CD34 image expression parameters are graphically represented for each group in scatterplots. For the analysis, data are expressed as mean ± SEM from 3 separate animals. Data for 0 h are represented by closed circles, 9 h by closed squares, and 24 h by closed triangular data points.

* P < 0.05. ** P < 0.01.

CD34 localizes in the canine bronchiolar epithelium, alveolar septa, and lamina propria

Lung sections from clinically normal dogs stained with rabbit anti-CD34 mAb showed that CD34 was strongly expressed in the cilia, surface, plasma membrane, and cytoplasm near the plasma membrane of the bronchial epithelium (Figure 4A), similar to what was observed in the clinically normal mouse lung. CD34 was also expressed in occasional cells in the lamina propria. Figure 4B showed that CD34 was present in the alveolar septum and the alveolar macrophage (inset). We also checked canine immunoreactivity of the CD34 antibody, which is primarily predicted for mouse reactivity, and SDS-PAGE revealed that the antibody raised against the CD34 C-terminal detected canine CD34 (Figure 4F).

Figure 4.

Canine lung immunohistochemistry for CD34. Representative lung sections from healthy normal dogs stained with rabbit monoclonal anti-CD34 (ab81289 clone EP373Y) (A and B). A — Strong CD34 expression (arrows) in the cilia, surface, plasma membrane, and cytoplasm near the plasma membrane of the bronchial (Br) epithelium. CD34 is also expressed in occasional cells in the lamina propria. B — CD34 in the alveolar (Av) septum and Av macrophage (inset). C — von Willebrand factor antibody reacting with the vascular endothelium on a blood vessel (BV) is a positive control (dark purple-pink) (n = 4). Magnification: 1000×. Lung sections from healthy normal dogs stained with rabbit IgG isotype control (D) or bovine serum albumin (E), serve as negative controls (methyl green with light background staining). F — Representative western blot performed on normal dog lung homogenates (n = 4). The red band corresponds to positive staining for CD34 and the green band is staining for β actin (control).

CD34 localizes in the porcine bronchial epithelium, the alveolar epithelium, and alveolar macrophages

In clinically normal pig lungs, CD34 was expressed in the bronchial epithelium (Figure 5A), cells of the alveolar septum (Figures 5A and B), and alveolar macrophages (Figure 5B, inset). In H3N1 influenza-infected lungs, CD34 was distributed mainly and prominently in the apical and basolateral surfaces of the bronchiolar epithelium (Figure 5C). Figure 5D shows intense staining on the surface of cells, likely pulmonary intravascular macrophages and other inflammatory cells, located in the alveolar septum. The vWF antibody stained the vascular endothelium (Figure 5E) while the omission of a primary antibody resulted in the lack of staining (Figure 5F).

Figure 5.

CD34 immunohistochemistry in porcine lungs. Immunohistochemical staining with rabbit monoclonal anti-CD34 (ab81289 clone EP373Y) was performed on normal pig lungs (A, B) and H3N1 influenza-infected pig lungs (C, D). In normal pig lungs, CD34 (arrows) is expressed in the whole body of the bronchial (Br) epithelium (A). CD34 is also present in type 1 and type 2 epithelium in alveolar (Av) septum (A, B) and alveolar macrophage (B, inset). In H3N1 influenza-infected lungs, CD34 (arrows) is distributed mainly on the apical side and around the border of the bronchial epithelium (C), type 1 and type 2 epithelia in the alveolar septum (D), alveolar macrophage (D, left inset), and neutrophil (D, right inset). A porcine lung section stained with anti-von Willebrand factor antibody (E), reacting with the vascular endothelium on a blood vessel (BV), is a representative of the positive control color for the protocol (dark purple-pink). Porcine lung sections stained with rabbit IgG isotype control (F) or bovine serum albumin (G) serve as negative controls (methyl green with light background staining) (n = 4).

Magnification: 1000×.

Discussion

A novel C-terminal targeted CD34 antibody was used in this CD34 antibody-based lung localization and expression study. The N-terminal or the extracellular CD34 region is heavily glycosylated and therefore, not appropriate for western blot quantification studies under reduced conditions. The cytoplasmic region on the other hand, comprises the conserved signaling component of the CD34 molecule. We validated the antibody for specificity against KO mice, which confirmed the feasibility of detailed staining experiments in normal and inflamed mice and pigs. We also checked for antibody cross-reactivity in canine species. Therefore, we provide a tool for the study of CD34 biology in mice, pigs, and dogs.

Our study has consistently and reliably shown the lung epithelial staining pattern of CD34 protein. Interestingly, 9 and 24 h after LPS exposure, CD34 expression became higher in brighter CD34 positive pixels (frequency distribution or skewness). At 9 h post-LPS exposure, CD34 protein expression was discrete or punctate compared to the diffuse pattern at 0 h (P < 0.06). Although CD34 kurtosis (i.e., spatial distribution) did not achieve statistical significance at 9 h (P = 0.06), it might still be biologically significant, considering the low number of mice used in our study. Overall, our results point towards putative cellular CD34 compartmentalization and probably internalization, although we did not use cell surface or organelle markers to explore this. Further studies will be required to uncover sub-cellular localization of CD34. Our observations of CD34 expression in the bronchiolar epithelium, together with the pan-selectin affinity of CD34 (17–20), reinforce a plausible role of CD34 in lung neutrophil adherence as indicated by higher lung MPO in LPS exposed WT mice when compared to CD34^−/− (data not shown).

Another major observation of the C-terminus targeting CD34 antibody is its cross-reactivity and the similar localization pattern in murine, canine, and porcine lungs. In earlier canine reports, CD34 expression has been observed in hematopoietic stem cells in bone marrow as well as peripheral blood (14), but scant to no information is available on canine lungs. We show bronchiolar and alveolar epithelial and alveolar macrophage localization of CD34 in both species. Remarkably, influenza-infected porcine lungs show additional staining in recruited leukocytes. The expression of CD34 in vascular and epithelial cells as well as immune cells suggests a potential role for this molecule in lung inflammation and inflammatory cell recruitment.

Much of the knowledge relies on N-terminal CD34 immunoreactivity, which identifies both epithelial type 1 and 2 precursors or bronchioalveolar stem cells and not bronchiolar epithelial cells per se (21). Our findings suggest a new tool to study CD34 localization and function. To our knowledge, these observations are reported for the first time in the lung tissues with the use of a novel antibody, EP373Y, which is reactive to the conserved C-terminal sequence of human CD34 protein.

Supplementary Data

Supplementary Figure 1