Apoptosis Inhibitor of Macrophages in Cats: A Potential Link Between an Exon 3 Variant Allele and Progression of Naturally Occurring Chronic Kidney Disease

2.1 Study Design

The study was carried out in four stages. First, DNA samples from cats collected from a study completed at Washington State University (WSU; protocol #4915) were screened to identify those with each fAIM exon 3 variant: homozygous wild-type, homozygous variant, or heterozygous (results not shown). Then, the fAIM gene in samples from cats identified as carrying 2, 3, and 4 copies of exon 3 were sequenced as described below. Second, medical records from client-owned cats admitted to the Veterinary Teaching Hospital at WSU were classified phenotypically following the inclusion and exclusion criteria described below (Figure 2). Third, archived (PrIMe DNA Bank-IACUC protocol #6493) genomic DNA from those cats meeting the inclusion criteria were classified as homozygous for the wild-type (2 copies of exon 3) and variant (3 and 4 copies of exon 3, respectively) alleles of the fAIM gene using droplet digital PCR (ddPCR). One of the researchers (NFV) determined the final list of CKD cases included in the study and remained blinded to the genotype. Lastly, a subset of archived genomic DNA from the PrIMe DNA Bank feline samples (IACUC protocol #6493) was genotyped using ddPCR to obtain preliminary information about the frequency of the variant in the cat population.

2.2 Genomic DNA Isolation

Genomic DNA from feline samples was isolated from cheek swabs using Quick-DNA miniprep kit (Zymo Research, Irvine, WA, USA) following the manufacturer's protocol. After extraction, the eluted DNA concentration was determined using Qubit fluorometric quantification following the manufacturers' instructions (Invitrogen, Carlsbad, CA, USA). The DNA samples were stored at −80°C until used.

2.3 Sequencing of Feline Apoptosis Inhibitor of Macrophage Gene (fAIM) Variants

Genomic DNA from archived feline samples (IACUC protocol #4915) was used for sequence studies. The complete DNA sequence of fAIM was amplified for PacBio sequencing, employing the following PCR parameters: a single cycle at 94°C for 30 s, followed by 30 cycles at 94°C for 30 s and 65°C for 750 s, with a final extension at 65°C for 10 min. In each PCR reaction, approximately 50 ng of genomic DNA template, 1× LongAmp Taq reaction buffer, 0.4 μM of each primer (Fa2 5′ CTGCAGACAAGACCTCTGTC 3′ and Ra1 5′ CAAGGCGTAGGCAGAGTTCAG 3′), 0.3 mM dNTPs, 5% dimethylsulfoxide (DMSO; Sigma–Aldrich, St. Louis, MO, USA), 2.5 U of LongAmp Taq DNA polymerase (New England Biosciences, Ipswich, MA, USA), and 10 μL of nuclease-free water were added for a final reaction volume of 25 μL.

Post-PCR amplification, the product was visualized on a 0.8% agarose/Tris acetate-EDTA (TAE) gel containing 0.5 μg/mL of ethidium bromide. Upon observing a PCR band of the expected size (approximately 11 kb), the PCR product was purified using 0.8× Ampure XP magnetic beads (Agencourt Biosciences, Beverly, MA, USA); product quality was assessed on the Agilent 5200 Fragment Analyzer (Agilent, Santa Clara, CA, USA) and subjected to PacBio sequencing at the WSU Laboratory for Biotechnology and Bioanalysis.

For the preparation of the PacBio SMRTbell library, the PCR products of 10–15 kb were employed by utilizing the Pacific Biosciences SMRTbell template preparation kit 1.0 (Pacific Biosciences, CA, USA) according to the manufacturer's instructions. A separate library was created for each amplicon, and purification was done using 0.8× Ampure XP magnetic beads (Agencourt Biosciences, Beverly, MA, USA). These purified libraries subsequently were sequenced using the P6C4 sequencing chemistry on the RS II platform, with a 240-min movie time and one SMRT cell per library.

The sequence analysis was carried out using the Pacific Biosciences SMRT Analysis software suite for long amplicon analysis and hierarchical genome assembly process (HGAP), resulting in polished amplicon sequences. Further sequence analysis was performed using Lasergene v. 14 (DNAstar, WI, USA).

2.4 Simulation of the Molecular Structure of 3-SRCR and 4-SRCR Domain Variants of fAIM and Estimation of Their Physicochemical Parameters

The nucleotide sequence for both the 3-SRCR and 4-SRCR domain variants of fAIM was converted to amino acid sequence using EditSeq in Lasergene v. 14 (DNAstar Madison, WI, USA). The corresponding amino acid sequences can be found in Supplemental information (Table S1). Initially, the amino acid sequence's signal was determined using deep neural networks as implemented through SignalP 5.0 (http://www.cbs.dtu.dk/services/SignalP/index.php) [14]. Utilizing this signal sequence information, the mature protein sequences and the predicted tertiary structures for both fAIM proteins were modeled using deep learning algorithms on the RaptorX server (http://raptorx.uchicago.edu/) [15, 16]. The predicted tertiary structure images were regenerated using UCSF Chimera v1.14 (https://www.cgl.ucsf.edu/chimera/download.html) [17, 18].

2.5 Case Selection

Medical records of cats presented to the Veterinary Teaching Hospital at WSU and diagnosed with CKD between 2011 and 2022 were screened for inclusion in the study (Figure 2). Records that indicated both a diagnosis of CKD and archived DNA were pre-selected. The following data were extracted from medical records: age at diagnosis of CKD, sex, reproductive status, breed, body weight, SCr, blood urea nitrogen concentration (BUN), urine specific gravity (USG), and serum thyroxine concentration (T4).

Cases were eligible for inclusion if they met the following criteria: (1) SCr ≥ 1.6 mg/dL together with a low USG (< 1.035) [19], or SCr ≥ 1.6 mg/dL on two occasions a minimum of 3 months apart [19], and (2) follow-up data for more than 6 months. The follow-up period was defined as the interval between the date of CKD diagnosis (designated baseline) and the date of the last available SCr (designated endpoint).

Cases were excluded if follow-up data was available for less than 6 months. Cats with other diagnosed renal conditions (e.g., ureteral obstruction, pyelonephritis, renal lymphoma, polycystic kidney disease, and subcutaneous ureteral bypass placement) were excluded. Cats with diagnosed hyperthyroidism were ineligible if T4 concentration was outside of the reference concentration range (1–4 μg/dL) during the follow-up period. Cats under treatment for hyperthyroidism with stable T4 concentrations within the reference concentration range (1–4 μg/dL) were not excluded. Follow-up data was not included if dehydration was reported. Serum creatinine concentration data from cats receiving fluid therapy or drugs suspected to alter SCr (e.g., doxorubicin, carboplatin, nonsteroidal anti-inflammatory drugs, gentamicin), or with a history compatible with an episode of acute disease in a patient with CKD, were excluded from the analysis.

2.6 Classification of Cases Based on Renal Function

Assessment of renal function was based on SCr reported in the medical records during the follow-up period. We used two approaches for phenotypic classification and analyzed them independently (Figure 2).

2.7 Determination of fAIM Exon 3 Copy Number

After assessment of kidney function, fAIM exon 3 copy number variation was determined using archived (PrIMe DNA Bank (IACUC protocol #6493)) genomic DNA samples from CKD cats that met the inclusion criteria.

Exon 3 copy number variation (CNV) was determined using ddPCR. Approximately 100 ng of DNA was combined with a reaction mixture composed of 2× Bio-Rad ddPCR mix and 20× primer and TaqMan probe mixes for fAIM exon 3 as the region of interest (HEX) and an internal (exon 5 of fAIM; exon 5) and external (exon 3 of feline telomerase reverse transcriptase; TERT) reference gene (FAM). Then, droplets were generated in a droplet generator (QX200, Bio-Rad, USA) by loading the reaction mixture containing the genomic DNA (gDNA) sample into DG8 reaction cartridges (Bio-Rad, USA) with 70 μL droplet generation oil (Bio-Rad, USA). After droplet generation, droplets were transferred to a 96-well PCR plate (Bio-Rad, USA). The PCR amplification was performed in a T100 Touch thermal deep-well thermocycler (Bio-Rad, USA) with the following parameters: 96°C for 10 min followed by 45 cycles of 96°C for 15 s and 60°C for 2 min, and finally 96°C for 10 min and a 4°C infinite hold, all with a 2°C/s ramp rate. The PCR plate was cooled to room temperature before loading the plate in the QX200 digital droplet reader (Bio-Rad, USA).

Fluorescence data were acquired and analyzed using QuantaSoft Analysis Pro Software version 1.05.596 (Bio-Rad, USA) following manufacturer recommendations. Copy number variation determination involves quantification of target and reference loci (TERT or exon 5 for our study). The ddPCR software counts positive and negative droplets for each of the targets (exon 3 in fAIM gene and each of the reference genes), estimating the average CNV of exon 3 per droplet. The software was set to calculate the copy number automatically based on input data. Each study sample was run in combination with both of the reference loci once.

2.8 Frequency of fAIM Exon 3 Genotypes From Archival DNA

The frequency of cats with 2, 3, and 4 copies of exon 3 of fAIM was determined using 100 archived DNA samples from the PrIMe DNA Bank. A subset of samples was selected from > 1500 banked samples by simple randomization. Genotyping for fAIM exon 3 CNV was determined using ddPCR as described above.

2.9 Statistical Analysis

Age at diagnosis, duration of follow-up period, SCr at baseline and endpoint, percentage change in SCr from baseline, BUN, and body weight for the cats were evaluated descriptively and compared statistically. Shapiro–Wilk and Brown–Forsythe tests were used to verify normality and the homogeneity of variances, respectively. An unpaired t-test was used to compare variables at the baseline or endpoint for cats within progressive and non-progressive groups. Mann–Whitney's test was performed if the data was not normally distributed. A paired t-test was used to compare body weight and SCr at baseline and endpoint within each group. Spearman's correlation coefficient was used to investigate the correlations between body weight and SCr. The strength of the association between fAIM exon 3 CNV and phenotype (IRIS stage progression and SCr worsening) was determined by calculating the odds ratio (OR) and tested statistically using Fisher's exact test. The OR and 95% confidence interval (CI) were calculated as described previously [20]. Statistical analyses were performed using Microsoft Excel and SigmaPlot version 14.0 (Systat Software Inc., California). For all tests, p ≤ 0.05 was considered significant.

3.1 Sequencing of Feline Apoptosis Inhibitor of Macrophage (fAIM) Gene

To ascertain the genomic status and provide a comprehensive understanding of exon 3 structure, three feline gDNA samples were subjected to long-read sequencing. These samples, previously identified with 2, 3, and 4 copies of exon 3 by qPCR screening (data not shown), underwent precise examination. A summary of the fragment analyzer run is included in the Supporting Information (Figure S1). Fragment analysis of the fAIM PCR product indicates three amplicon sizes between approximately 10 and 12.5 kb pairs (Figure S2). In the sample with approximately 10.4 kb pairs, each allele contained an exon 3 of 321 base pairs, indicating a homozygous wild-type gene (Figure S2, Panel A). For the sample with approximately 12.5 kb pairs amplicon, both alleles contained 2 copies of exon 3 of 321 base pairs, reflecting a homozygous state for the mutant allele (Figure S2, Panel B). The sample with both size amplicons contains an allele with an exon 3 and the other allele with 2 copies of exon 3, indicative of a heterozygous state for the mutant allele (Figure S2, Panel C). Exon 3 copies exhibited a homology range of 96.2%–100%. The altered fAIM gene consists of a large novel intragenic tandem exon 3 and flanking intron duplication of the fAIM gene (Figure S2, Panel A, the intron size is approximately 2.5 kb pairs). The wild-type fAIM gene is composed of six exons (exons 1–6; Figure S2, Panel A). Detailed maps of fAIM introns and exons for both wild-type and mutant alleles are provided in Figure S2, Panel B.

3.2 Simulation of the Molecular Structure of 3-SRCR and 4-SRCR Domain Variants of fAIM and Estimation of Their Physicochemical Parameters

The copy number variation of feline exon 3 results in an additional SRCR domain 1 in the protein structure (Figure 3A,B). Thus, affected cats express a 4-SRCR domain variant of fAIM. To explore the potential impact of the extra protein domain on the protein's physicochemical features, we modeled in silico the molecular structure of 3-SRCR and 4-SRCR domain variants of fAIM and estimated their physicochemical parameters. A protein simulation model for both protein variants is shown in Figure 3, Panel A and B. Amino acid sequences for 3-SRCR and 4-SRCR domain fAIM variants are presented in Table S1. The 4-SRCR domain protein variant results in a duplication of the first domain, resulting in a larger, more negatively charged, and less water-soluble protein than the wild-type 3-SRCR domain fAIM protein (Table S2).

3.3 Phenotypic Case Selection

Initially, 172 medical records of cats diagnosed with CKD during the designated time period were identified. Ninety-one of these also had archived DNA, and thus, only those medical records were reviewed. Fifty cases met the inclusion criteria (Table S3). Twenty cases had SCr ≥ 1.6 mg/dL and USG < 1.035 at diagnosis, and 30 had SCr ≥ 1.6 mg/dL at least twice ≥ 3 months apart. Forty-one cases were excluded because follow-up data duration was < 6 months (n = 24), other renal conditions were present (n = 6), T4 concentrations were abnormal (n = 6) or clinical information was incomplete (n = 5, Table S4).

The most common breeds represented were domestic short hair (29/50; 58%), domestic long hair (7/50; 14%), and Siamese or Siamese-mix (6/50; 12%). Other breeds were Bengal (2/50; 4%), Ragdoll (2/50; 4%), Bombay (1/50; 2%), domestic medium hair (1/50; 2%), Egyptian Mau (1/50; 2%), and Persian (1/50; 2%; Table S3). Of 50 cats, 28 (56%) were neutered males, 21 (42%) were spayed females, and 1 (2%) was an intact male (Table 1). Age of cats at CKD diagnosis ranged from 8 to 15 years for all cats, and no significant difference was found in age at diagnosis between cats with progressively increased SCr versus those that remained non-progressive (p = 0.90, Table 1).

3.4 Classification of Patients According to Renal Function

3.4.1 Approach A: IRIS Staging

Cats were classified as IRIS stage 2 (n = 43) and stage 3 (n = 7) at baseline. Most of the cats (n = 29) remained in the same IRIS stage at the end of the follow-up period. Twelve (24%) cats progressed from stage 2 to 3 and two cats from stage 2 to 4. Four cats changed from IRIS stage 2 to 1 and two cats from IRIS stage 3 to 2 (Figure 4A and Table S3).

3.5 Number of Copies of fAIM Exon 3

Fifteen of 50 (30%) cats were homozygous for the wild-type allele of fAIM, and 35 of 50 (70%) cats were variant-type (> 2 copies of exon 3, Table S3). Within the fAIM variant group, 26 of 50 (52%) cats were heterozygous for the variant allele of fAIM (3 copies of exon 3), whereas 9 (18%) cats were homozygous for the variant allele of fAIM (4 copies of exon 3, Table S3).

3.6 Association Between Copies of fAIM Exon 3 and Change in SCr

A significant association was found between fAIM exon 3 copy number variation and IRIS staging change (Figure 4, Panels A and B, Table S2). Cats with 4 copies of exon 3 of fAIM had higher odds of worsening (p = 0.01) IRIS stage (approach A) and SCr than cats with wild-type fAIM exon 3 (approach B, p = 0.03, Table 2).

3.7 Frequency of fAIM Exon 3 Genotypes in a Subset of DNA Samples From WSU DNA Bank

Of the selected 100 banked feline DNA samples, 38% were homozygous for the wild-type fAIM exon 3 allele, 20% of the cats were homozygous for the variant fAIM exon 3 allele, and the remaining 42% were heterozygous.