High Incidence of Sebaceous Gland Inflammation in Aldose Reductase-Deficient Mice

Abstract

Aldose reductase is a member of the 1B1 subfamily of aldo-keto reductase gene superfamily. The action of aldose reductase (AR) has been implicated in the pathogenesis of a variety of disease states, most notably complications of diabetes mellitus including neuropathy, retinopathy, nephropathy, and cataracts. To explore for mechanistic roles for AR in disease pathogenesis, we established mutant strains produced using Crispr-Cas9 to inactivate the AKR1B3 gene in C57BL6 mice. Phenotyping AR-knock out (ARKO) strains confirmed previous reports of reduced accumulation of tissue sorbitol levels. Lens epithelial cells in ARKO mice showed markedly reduced epithelial-to-mesenchymal transition following lens extraction in a surgical model of cataract and posterior capsule opacification. A previously unreported phenotype of preputial sebaceous gland swelling was observed frequently in male ARKO mice homozygous for the mutant AKR1B3 allele. This condition, which was shown to be accompanied by infiltration of proinflammatory CD3+ lymphocytes, was not observed in WT mice or mice heterozygous for the mutant allele. Despite this condition, reproductive fitness of the ARKO strain was indistinguishable from WT mice housed under identical conditions. These studies establish the utility of a new strain of AKR1B3-null mice created to support mechanistic studies of cataract and diabetic eye disease.

1. Introduction

Diabetes mellitus is recognized as a leading cause of new cases of blindness in patients between the ages of 20 and 74 [1]. In addition to retinopathy, diabetic patients are also at a much higher risk for corneal defects [2] and cataracts [3, 4]. These findings reaffirm that the pathogenesis of diabetic eye disease is complex and likely involves many intersecting pathways of metabolism and gene regulation. A tremendous amount of work has gone into elucidating molecular mechanisms behind diabetic blindness. Theories include excess formation of advanced glycation end-products (AGEs), activation of the glucosamine pathway, activation of PKC isoforms, and activation of the polyol pathway [5]. Unfortunately, treatment of diabetic patients with drugs targeting pathways associated with each of these disease mechanisms has not been successful. An alternative approach to discovery of novel therapeutic strategies is to disrupt metabolic and signaling pathways through creation of mouse strains with mutations that result in loss of function of targeted gene products. So-called knock out mice have been invaluable in providing new mechanistic insights into complex disease phenotypes in virtually all major organ and sensory systems. For example, enhancement of AR activity by creating transgenic animals causes exacerbation of diabetic eye disease, including cataract [6, 7] and retinopathy [8, 9]. Conversely, inactivation of the AR gene protects against diabetes-induced cataract and histopathological markers of retinopathy such as pericyte loss, blood-retinal barrier breakdown, increased VEGF, and markers of retinal nitrosative stress [10, 11].

In the course of our studies of diabetic eye disease, we produced two new strains of AR knock out (ARKO) mice by Crispr-Cas9 gene disruption. Through careful phenotyping of the new strains, we observed a relatively high frequency of swollen preputial glands in male ARKO mice. Preputial glands are sebaceous glands that produce lipid and hormone secretions that are thought to contribute to olfactory signaling among animals. Histological studies demonstrated that swollen glandular tissue contained elevated levels of inflammatory cells in granulomas, implicating loss of AR with changes in regulatory signaling that would normally control inflammation and infiltration of inflammatory lymphocytes. This constitutes the first report of loss of AR being associated with sebaceous gland abnormalities in an animal model.

2. Materials and Methods

2.1. Generation of AKR1B3 null mice (ARKO).

The AKR1B3 knock out strain in C57BL6 mice (ARKO) was generated with CRISPR/Cas9 technique using the assistance of the Transgenic Core of the Gates Center for Regenerative Medicine at the University of Colorado School of Medicine |Anschutz Medical Campus. A single guide RNA (sgRNAs; 5′-CCCCTCACCTTGAACCCCGT-3′) was designed to target exon 3 of the Akr1b3 gene. Cas9 mRNA and sgRNA were electroporated into one cell stage zygotes collected from C57BL/6 mice, which were then subsequently implanted into the oviducts of pseudopregnant mice. Neonatal mutant mice were identified by genotyping and sequencing. Expected rates of transmission of the mutant AKR1B3 allele were observed following backcrossing to C57BL/6 for 5 generations. Thereafter, homozygotes for the mutant allele were produced through heterozygous matings, and confirmed by observing the expected rates of heterozygotes for the mutant allele produced by backcrossing to WT mice.

2.2. Enzyme and metabolite determinations.

6 male and 6 female ARKO mice, and 3 male and 2 female AR transgenic mice (AR-Tg)[12] were euthanized with CO₂ and both kidneys surgically removed. The kidneys were weighed and homogenized in phosphate buffered saline (PBS) using a ratio of 80 mg kidney tissue to 200 μL PBS with a handheld tissue homogenizer. Homogenates were then centrifuged at 4°C, 14,000×g for 5 min to generate a pellet plus the aqueous (lower) and fatty (upper) layers. The aqueous layer was removed by pipet as the final homogenate and stored at −80°C.

AR activity was measured at two levels. First, the aldo-keto reductase activity of AR was measured directly in tissue homogenates in a microplate format (ab273276 from Abcam, Waltham, MA, USA) to measure AR-mediated NADPH oxidation in the presence of DL-glyceraldehyde. For a second method to ascertain AR activity in mouse tissues, we measured tissue levels of sorbitol to assess the functional consequence of AR gene silencing. Sorbitol levels were measured using a microplate-based assay (EnzyChrom^™ Sorbitol Assay Kit (ESBT-100), BioAssay Systems, Hayward, CA, USA). Homogenates were not filter deproteinated prior to quantification. Homogenates were diluted 10-fold with water prior to assaying. Water was used as the sample blank (i.e.: 0 μM sorbitol standard) and sorbitol concentrations were calculated as indicated by the manufacturer after reading the plate on a SpectraMax iD3 (Molecular Devices, Sunnyvale, CA, USA) at 565nm, with a standard curve ranging from 0–1000 μM.

After enzyme and metabolic assays, protein content in samples was determined using a bicinchoninic acid (BCA) assay (Pierce^™ BCA Protein Assay Kit, Pierce Biotechnology, Rockford, IL, USA) with albumin as a standard ranging from 0 to 2000μg/mL. Homogenate samples were diluted 25-fold before assay. Data were analyzed using in GraphPad Prism 9.4.1. Values from ARKO and AR-Tg mice were compared using an unpaired two-sample t-test assuming equal variances, supported by an F-test.

2.2. Immunohistochemistry and Western Blot Analysis

Immunohistochemistry was performed on paraffin sections after antigen retrieval with 6 minutes 110°C in sodium citrate buffer (Sigma, St Louis, MO) then 10 minutes in 3% hydrogen peroxide and blocked for 20 minutes with 2.5% horse serum (Vector Laboratories, Newark, CA). CD3 was detected using 25-fold diluted anti-CD3 antibody (Dako, Glastrup, Denmark), incubated for 1hr, followed with anti-rabbit HRP and visualized after developing with diaminobenzidine (Vector Laboratories, Newark, CA). AR was detected using 200-fold diluted anti-aldose reductase (ab153897, Abcam, Cambridge, UK) incubated for 1hr followed with anti-rabbit HRP (Vector Laboratories, Newark, CA) and visualized using diaminobenzidine staining (DAB peroxidase, Vector Laboratories, Newark, CA).

For Western Blot analylsis, tissue homogenates were cleared by centrifugation, with the resulting supernatants (1μg of as determined by BCA) resolved by TGX gel electrophoresis (Biorad, Hercules, CA). and transferred to PVDF membrane. Membranes were blocked with 1% milk in TBST and incubated with 1:1000 anti-aldose reductase antibody (ab153897, Abcam, Cambridge, UK) or β-actin (sc47778, Santa Cruz, Dallas, TX). Immune complexes were visualized after treatment with 1:1000 secondary antibody (Jackson Laboratories, West Grove, PA). Chemiluminescence was performed using Super Signal (Peirce, Waltham, MA) on a Chemidoc Imaging System (Biorad, Hercules, CA).

2.3. Extracapsular Lens Extraction (ECLE)

Extracellular lens extraction (ECLE) was performed as previously described [13]. Lens capsules were collected 5 day post-surgery and placed into 500μL of Qiazol (Qiagen, Austin, TX) and 100μL chloroform (Sigma, St Louis, MO) and frozen to −80°C. RNA was extracted using micro RNeasy (Qiagen, Austin, TX). Preparation of cDNA was carried out with iScript Reverse Transcription Supermix for RT-qPCR (Biorad, Hercules, CA) and qRT-PCR was performed using iTaq Universal SYBR Green Supermix (Biorad, Hercules, CA) on a CFX Connect (Biorad, Hercules, CA) using prequalified primers (Integrated DNA Technologies, Coralville, IA). PCR primers were as follows: αSMA forward 5′-CTGTTATAGGTGGTTTCGTGGA-3′ reverse 5′-GAGCTACGAACTGCCTGAC-3′, GAPDH forward 5′-AATGGTGAAGGTCGGTGTG-3′ and reverse 5′-GTGGAGTCATACTGGAACATGTAG-3′. Relative fold-expression was determined using the 2-ΔΔCT method on N=4 in triplicate. Statistical analysis was performed using Graphpad Prism (Graphpad Software, La Jolla, CA, USA). Data was analyzed by ANOVA with Tukey’s post hoc test. Asterisks refer to P values according to *P<0.05, **P<0.01, and ***P<0.001.

3. Results

3.1. Generation and characterization of AKR1B3 null mice

Two independent founder strains were identified in mice generated by the gene targeting experiments. Sequencing of the AKR1B3 locus revealed one mutant with a 16 bp deletion in exon 3 (named ARKO-16), while another revealed a 4 bp insertion at exon 3 (named ARKO+4). In both cases, Crispr-Cas9-mediated mutations are predicted to give rise to transcripts containing an in-frame stop codon following the mutation site. (Supplementary Figure 1). Western blot analysis tissue extracts from WT and mutant lines confirmed the loss of AR expression in mice homozygous for the mutant allele (Fig 1B). The probe used in these experiments was a monoclonal antibody specific for an epitope mapping between amino acids 115–140 (sc-166918 Santa Cruz Biotechnology, Inc.). As this epitope domain would be encoded immediately downstream from the mutation site, we predicted these mutations would lead to loss of immunoreactivity for AR in both strains. Accordingly, we carried out Western blot analysis of lens extracts produced from WT and mutant mouse tissues. As shown in Fig 1A, AR expression was readily detected in epithelial cell layers of the cornea and lens from WT mice but was essentially undetectable in the corresponding tissues in ARKO mice (ARKO-16). Similarly, AR was easily measured by Western blotting from WT lens homogenates but was not detectable in lens extracts from ARKO strains (Fig 1B). In concordance with immunochemical results, we observed significant reductions in enzyme activity levels when tissue extracts from ARKO and WT animals were assayed for NADPH-dependent AR activity. Using DL-glyceraldehyde as the carbonyl substrate, kidney extracts from AR-Tg mice showed 33.4±4.4 mU/mg AR activity, WT had 24.2±8.5 mU/mg, and ARKO extracts showed 1±0.9 mU/mg (Fig 1C). In addition, we compared AR-Tg and ARKO tissue for levels of sorbitol, the in vivo product of AR catalytic conversion of glucose to sorbitol. As shown in Figure 1D, sorbitol levels in kidney extract were reduced approximately 2.5-fold in ARKO as compared to AR-Tg tissues. We consider it likely that basal levels of sorbitol did not completely diminish in the ARKO kidney, since sorbitol dehydrogenase can catalyze the NADH-dependent conversion of fructose to sorbitol and thus backfill reduced levels of sorbitol in the ARKO kidney [14, 15].

Figure 1.

Validation of AR gene silencing in ARKO mice. A. Immunohistochemical staining for AR in ocular anterior segments of WT and ARKO mice. Arrows point to epithelial layers of the cornea and lens. Separation of the lens and cornea layers occurred during histological processing of the ARKO sample and is not considered a phenotype of the knock out strain. B. Western blots of lens homogenates from WT and ARKO strains showing loss of AR immunoreactivity in ARKO tissues. β-actin is included as a loading control; C. AR enzymatic activity in kidney extracts from AR-Tg, WT, and ARKOmice, * p<0.01; D. Sorbitol levels in AR-Tg and ARKO kidney tissue, **p<0.005. Animals were 20–25 weeks of age. ARKO+4 strain was used in these measurements.

3.2. Ocular phenotypes in ARKO mice

Based on gross appearance, as well as by the organization of the usual layers of specialized cells in the eye (cornea, ciliary body, lens, retina), there were no noticeable differences between eyes of WT and ARKO mice (Fig 2). Immunostaining for AR, which was robust in the epithelial layers of the cornea and lens of WT mice, was markedly decreased in the corresponding tissue layers in the ARKO strain (Fig 1). AR expression in lens epithelial cells has been shown to influence the severity of epithelial-to-mesenchymal transition in the eye following cataract extraction [13, 16]. As shown in Fig 3, lenses from the ARKO+16 had a reduced EMT response, in comparison with AR-Tg control mice, as exemplified by a reduced levels of αSMA expression measured 10 days following extracapsular lens extraction. Previous studies showed that WT and AR-Tg mice have a quantitatively-similar response in αSMA gene transcript levels following lens extraction [13]. In concordance with previous studies, E-cadherin transcript levels, which dropped dramatically in AR-Tg lens capsules following lens extraction, were largely preserved in ARKO mice (Fig 3).

Figure 2.

Normal eye morphology in AKR1B3 mutant mice. (A) Wild type; (B) AR-Tg; (C) ARKO strain +4; (D) ARKO strain −16. Histological sections from eyes of wild type and mutant mice were collected at 8 months of age and stained with hematoxylin an eosin. Scale bar is 500 μm. Major structures of the eye (cornea, lens, and retina) highlighted by arrows shown for a wild type mouse (panel A) appear to be largely undisturbed in the eyes of AR-Tg and ARKO mice shown in panels B, C, and D.

Figure 3.

Loss of EMT response following lens extraction in ARKO mice. The abundance of targeted gene transcripts was measured in cells associated with the lens capsule dissected from AR-Tg and ARKO mice 5 days after extracapsular lens extraction (ECLE). (A) αSMA gene transcripts; (B) E-cadherin transcripts. N ≥ 4 mice in each group; * p<0.01; ** p<0.001 established by ANOVA with Tukey’s post hoc testing (*). Animals were 20–25 weeks of age.

3.3. Preputial lesions in ARKO male mice

During the routine care and maintenance of our ARKO mice, we were frequently alerted by animal care staff to the development of a raised tissue mass in the genital area of a large proportion of male ARKO mice. Gross dissections revealed the raised tissue to be abscesses of the preputial gland (Fig. 4), which is an excretory tissue prone to bacterial infection and abscessation [17]. Immunohistochemical examination of the abscessed tissue from ARKO mice following staining with antibodies to CD3 revealed the presence of large numbers of inflammatory T-cells (Fig. 4); tissue taken from WT mice showed markedly fewer CD3-positive cells. When followed up to 12 weeks of age, preputial gland enlargement was observed in approximately 25% of male mice, and increased to 58% of male mice over time. The tendency toward development of a preputial mass seemed not to affect reproductive fitness of ARKO male mice. Litters sired by male breeders homozygous for the null AR allele comprised 5.3±1.5 (measured over 13 litters of ARKO-16 and 11 litters of ARKO+4) pups per litter, compared with 5.5±1.4 pups/letter produced by breeders with the WT AKR1B3 genotype (>20 litters of WT).

Figure 4.

Preputial gland abnormalities in ARKO mice. Gross appearance of preputial glands in WT (A) and ARKO (B) mice. Areas of interest are demarcated by red circle. Preputial tissues were dissected from ARKO and WT male mice and immunohistochemically stained for CD3+ cells using visualization with diaminobenzidine (brown) and counter-stained with hematoxylin. Representative immunopositive cells are indicated with arrows. (C) Wild type, (D) ARKO, (E) ARKO no primary antibody control. Scale bar is 50μm. Animals were 20–25 weeks of age.

4. Discussion

CRISPR-Cas9 has revolutionized the ability to initiate the gain and loss of function of specific enzymes in knockout mice in an efficient and simplistic manner [18]. This innovation has led to many significant recent discoveries, from cancer therapy utilizing CRISPR-Cas9 to confirm and discover new drug targets for the creation of anticancer drugs, to the modeling of eye diseases such as autosomal optic atrophy. [19, 20]. In the current study, we used a CRISPR-Cas9 strategy to produce mutant mice deficient in expression of aldose reductase, an aldo-keto reductase responsible for the NADPH-dependent reduction of glucose into sorbitol. Studies have implicated a role for AR in the pathogenesis of complications associated with diabetes mellitus, including cataract, retinopathy, neuropathy, and nephropathy [21]. While an AR-deficient mouse strain was previously generated using homologous recombination to disrupt the AR gene [22], continuous inbreeding of the strain over the ensuing over 2 decades could have led to the accumulation of new mutations that could influence phenotypes associated with retina development and diabetic eye disease. We therefore undertook an effort to produce a novel set of AR null mutants using CRISPR-Cas9 technology.

Two founder strains deficient in AR biosynthesis were identified in our studies, both containing mutations targeted to exon 3 of the mouse AR gene (AKR1B3). Strain ARKO-4 contained a deletion of 4 bp, whereas ARKO+16 contained a 16 bp addition. In both cases, mutations predicted a premature translational stop in the coding sequence downstream from the mutation sites, respectively. Mutations in both strains were transmitted in a predicted Mendelian fashion over 6 matings to WT breeders. There were no differences in fertility when comparing litter sizes from WT and ARKO mice.

Disrupted production of AR was confirmed by measuring for loss of enzyme at three different levels in ARKO strains. Loss of AR enzyme protein in the kidney, where AR is abundantly expressed in its role in production of sorbitol as an active osmolyte [23], was confirmed by Western blotting. Reduction in presumed AR activity was confirmed in ARKO strains through assays of NADPH-dependent reduction of DL-glyceraldehyde, and by significant reductions in sorbitol levels in kidney homogenates. In concordance with our published works using the previous ARKO strain to demonstrate a role for AR in epithelial-to-mesenchymal transition [13] of lens epithelial cells following lens extraction surgery, we observed a deficit in markers of EMT expressed in lens cells from our new ARKO strains following extracapsular lens extraction. Thus, the new ARKO strains faithfully reproduces this important phenotype and thus appears to be validated for future studies of lens fibrosis.

One previously unreported phenotype observed in ARKO strains was the frequent appearance of swelling associated with preputial glands in male mice. Preputial glands are lipid and hormone-secreting glands, unique to rodent species, that may contribute to olfactory signaling among animals [24, 25]. Histological sectioning shows that the tissue contains large cavernous ducts within a connective tissue capsule; ducts feed into a central conduit that runs the length of the gland. Several different conditions may lead to preputial swellings, including bacterial infection as well as ductal ectasia, hyperkeratosis, cyst formation, and a variety of proliferative and neoplastic changes that can develop with age [17, 26]. As in our current study, swelling of preputial glands has been reported as an unexpected phenotype in several mouse strains with transgenic and gene-targeted mutations unrelated to AR [27]. Of the reported preputial gland phenotypes relevant to the present study, the observation of enlarged glands in a Prdm1 knock out strain [28] and in mice deficient for Nucling, a novel apoptosis-associated molecule acting through galetin-3 [29], appear to share similarities in preputial gland abnormalities observed in our ARKO strain. Future studies will be required to test for functional connections that tie together pathogenesis of preputial gland abnormalities among these mutant strains.

5. Conclusions

The authors describe the generation and characterization of mouse strains carrying mutations to disrupt the expression of AKR1B3, the gene encoding aldose reductase. Such strains will be valuable resources to the research community to aid in studies of aldose reductase and its role in various signaling and metabolic pathways in health and disease.

Highlights.

• AKR1B3-null mice (ARKO) were produced by Crispr-Cas9-mediated gene disruption

• ARKO mice demonstrate reduced EMT response following lens extraction

• ARKO mice are at increased risk for development of preputial gland inflammation