Characterization of the cDNA Encoding αIIb and β3 in Normal Horses and Two Horses with Glanzmann Thrombasthenia

Abstract

Glanzmann thrombasthenia (GT) is an inherited, intrinsic platelet defect characterized by a quantitative or qualitative change in the platelet glycoprotein complex IIb-IIIa (integrin α_IIbβ₃). The subunits are encoded by separate genes and both subunits must be expressed for a stable complex to form on the platelet surface; therefore, a defect in either gene can result in GT.

Dr. Eduard Glanzmann, a Swiss pediatrician, first described Glanzmann thrombasthenia (GT) in his patients in 1918; however, the association of a deficiency of the platelet glycoprotein complex IIb–IIIa with GT was not made until 1974. ^{14– 16} In the decade that followed, this glycoprotein complex was recognized as being the receptor that mediated platelet aggregation and was termed the fibrinogen receptor. In the 1990s, it was recognized that the fibrinogen receptor of platelets was an integrin and was designated as α_IIbβ₃ to comply with the nomenclature for that system. It was recognized at this time that the subunits were encoded by separate genes and that both subunits were required for a stable complex to form on the platelet surface. This disease has been well documented at the clinical and molecular levels in human beings and dogs ^{2, 3, 7, 13} (http://sinaicentral.mssm.edu/intranet/research/glanzmann/menu). The first description of GT at the molecular level in humans was published in 1990, while the molecular basis for GT in dogs was first reported in 1999. ^{4, 6}

GT was recently described in two horses with histories of chronic, intermittent epistaxis that was unrelated to exercise. ¹² Horse No. 1 was a 7-year-old thoroughbred cross gelding located in Hatfield, Hertfordshire, England, while horse No. 2 was a 4-year-old quarter horse mare located in Auburn, Alabama. These animals had normal coagulation screening test results, normal platelet numbers, and normal von Willebrand factor antigen levels. Clot retraction and platelet aggregation responses were markedly impaired. Flow cytometric studies indicated a reduction in the α_IIbβ₃ integrin on platelet surfaces. Based on these findings, both horses were diagnosed with GT.

The purpose of this study was to determine the cDNA sequences encoding normal equine α_IIb and β₃ and compare them with established human and canine cDNA sequences and to cDNA sequences determined in both affected horses.

Platelet-rich plasma was obtained by centrifuging 100–200 ml of EDTA-anticoagulated whole blood as previously described. ⁵ Platelets were counted and concentrated into pellets containing 5 × 10⁹ platelets per pellet. Total RNA was harvested from platelet pellets using the Micro to Midi Total RNA Purification System (Invitrogen Corporation, Carlsbad, CA). First strand cDNA synthesis was performed using the Superscript™ First-Strand Synthesis System for reverse transcription-polymerase chain reaction (RT-PCR) using oligo(dT) primers (Invitrogen Corporation).

Genomic DNA was harvested from EDTA-anticoagulated whole blood using the QIAamp® DNA Blood Mini Kit (Qiagen, Inc., Valencia, CA). Most primers were designed based on homologous regions of canine (GenBank accession nos. AF153316 and AF116270) and human DNA sequences for α_IIb and β₃. ^{8, 17} Normal equine cDNA sequence was used to design some of the primers.

DNA segments were amplified by PCR in an overlapping fashion using normal equine cDNA or genomic DNA as templates initially (Tables 1 and 2). The experiments were then repeated using cDNA or genomic DNA isolated from horses 1 and 2. In selected experiments, genomic DNA from relatives of horse 2 (dam, sire, and two half-siblings) were also evaluated. Amplification products were separated via electrophoresis on 1.5% agarose gels. DNA was extracted from target bands using the QIAquick Gel Extraction Kit (Qiagen, Inc.). Harvested bands were sequenced directly by a laboratory using an ABI 3100 Genetic Analyzer. Nucleotide and amino acid sequences from different species were compared using the AlignX program of the VectorNTI suite (Informax Invitrogen Life Science Software, Frederick, MD).

Table. 1.

Primers and polymerase chain reaction (PCR) conditions used to amplify equine α_IIb cDNA.∗

Primer Name	Primer Sequence 5′ to 3′	PCR Conditions†	Target Size (bp)
PreEx 1 f	ATTCCTGCCTGGGAGGTTGTG	30 sec, 30 sec, 45 sec	634
Exon 5 r	GCCGGCTTCGCAGTAGCGCTT	55°C anneal, 40 cycles
Exon 4 f	ACGCCCGTAGGTGGCTGCTTT	30 sec, 30 sec, 45 sec	741
Exon 12 r	GTAGCCATCTCGGTTGAGGTC	55°C anneal, 40 cycles
Exon 12 f	CTGACTGGCACACAGCTCTAT	30 sec, 30 sec, 45 sec	372
Exon 15 r	CACAGCAGGATTCAGCGAATC	50°C anneal, 40 cycles
Exon 14 f	CTACTGGTAGGAGCTTATGGG	30 sec, 30 sec, 45 sec	764
Exon 21 r	CCAGCTCACACAGCACTATCT	45°C anneal, 40 cycles
Exon 21 f	GAGGCTCATCTGTGACCAGAA	30 sec, 30 sec, 45 sec	903
Exon 29 r	CAGCACCCACCAGATGGGAAT	53°C anneal, 30 cycles
Exon 28 f	ACAGTCACGCGCTTGGTTCAA	30 sec, 30 sec, 45 sec	285
UTR r	AGAATAGTGTAGGCTGCACCA	51°C anneal, 30 cycles

∗bp = base pairs; f = forward; r = reverse; sec = seconds; UTR = untranslated region.

†Denaturing, annealing, and extension times, respectively.

Table. 2.

Primers and polymerase chain reaction (PCR) conditions used to amplify equine β ₃ cDNA.∗

Primer Name	Primer Sequence 5′ to 3′	PCR Conditions†	Target Size (bp)
PreEx A f	CCGCGGGAGGCGGACGAGAT	30 sec, 30 sec, 45 sec	360
Ex B r	CTGGCCGGAGCCGGAGTGCAA	58°C anneal, 40 cycles
Ex A f	GGCCCAACATCTGTACCACGC	30 sec, 30 sec, 45 sec	510
Exon C r	GGCCTCTGGTGGGGAGATAAA	54°C anneal, 40 cycles
Exon C f	GCTTTTGTGGACAAGCCTGTG	30 sec, 30 sec, 45 sec	440
Exon F r	CTGGGAGAGCTTCTCTGTCAT	51°C anneal, 35 cycles
Exon D f	AAGCAGAGTGTGTCACGGAAC	30 sec, 30 sec, 45 sec	447
Exon H r	CAGCTCCACTTTAGAGCGGAT	52°C anneal, 40 cycles
Exon G f	CTCATCCCCGGAACCACAGTG	30 sec, 30 sec, 45 sec	642
Exon I r	CTGAGCACATCTCCCCCTTGTA	55°C anneal, 30 cycles
Exon I f	CCATCAAGCCTGTGGGCTTCA	30 sec, 30 sec, 30 sec	446
Exon J r	GCGCGTGGTACAGTTGCAGTA	55°C anneal, 30 cycles
Exon I2 f	TACAAGGGGGAGATGTGCTCA	30 sec, 30 sec, 45 sec	505
Exon M r	AAAGCAGGACCACCAGGATGT	51°C anneal, 40 cycles
Exon K f	GCCCTCTACGAGGAGAATACC	30 sec, 30 sec, 45 sec	420
Exon N r	AGTGCCCCGGTAGGTGATGTT	51°C anneal, 35 cycles
Exon N f	GCCACATCCACCTTCACCAAC	30 sec, 30 sec, 30 sec	429
UTR r	CAGCCCTGCCCTTCTCTTCAG	56°C anneal, 40 cycles

∗bp = base pairs; f = forward; r = reverse; sec = seconds; UTR = untranslated region.

†Denaturing, annealing, and extension times, respectively.

Normal equine cDNA sequence encoding α_IIb had 82% nucleotide identity with human and 87% with canine sequences. At the amino acid level, percent identities were 82 and 81%, respectively. The cDNA encoding β₃ in horses had 92% nucleotide identity with human and 93% with canine sequences. Amino acid identity for β₃ was 95% when compared with both human and canine sequences. Within the four calcium-binding domains of α_IIb, equine cDNA exhibited 86–88% identity with human nucleotide sequence. Within the four calcium-binding domains, amino acid identity between equine and human ranged from 83 to 91%. When equine nucleotide sequences were compared with canine sequences within the four calcium-binding domains, nucleotide identities ranged from 83 to 91%, and amino acid identities ranged from 75 to 100%. cDNA sequences for equine α_IIb and β₃ have been submitted to GenBank (GenBank accession nos. AY322154 and AY999064).

cDNA sequences encoding β₃ in horses 1 and 2 with GT were identical to sequences obtained from a normal horse. cDNA sequences encoding α_IIb were also identical except for a single guanine to cytosine (CGG to CCG) substitution in codon 41 in exon 2 (Fig. 1A, B). This change would result in the substitution of a proline for an arginine in a highly conserved region of the encoded protein.

Fig. 1.

Fig. 1A. cDNA sequence encoding beginning of exon 2 of α_IIb from a normal horse. The first arrow indicates the beginning of the first complete codon of exon 2. The second arrow indicates the second nucleotide position in codon 41. Fig. 1B. cDNA sequence encoding beginning of exon 2 of α_IIb in horse 2. The first arrow indicates the beginning of the first complete codon of exon 2. The second arrow indicates change of G to C at the second nucleotide position in codon 41. Horse 1 cDNA sequence was identical to horse 2 at this location. Fig. 1C. Genomic DNA sequence from horse 1. The first arrow shows the beginning of the first complete codon of exon 2. The second arrow indicates the nucleotide change of G to C in exon 2 of the α_IIb gene. This horse was homozygous for the cytosine for guanine substitution. Fig. 1D. Genomic DNA sequence from horse 2. The first arrow indicates the beginning of the first complete codon of exon 2. The second arrow indicates the presence of both C and G nucleotides (designated as N) in exon 2 of the α_IIb gene. This horse and the sire were heterozygous at this location. The dam and two half-siblings were homozygous for the normal sequence and phenotypically normal.

Exon 2 was evaluated by PCR amplification and sequencing using as template genomic DNA from 10 normal horses, from GT horses 1 and 2, and from the sire, dam, and two half-siblings of horse 2 that shared the same dam. Genomic DNA sequencing results indicated that horse 1 was homozygous for the single guanine to cytosine base change, while horse 2 was heterozygous for the base change (Fig. 1C, D). The genomic DNA sequences encoding exon 2 of 10 normal horses and the dam and two siblings of horse 2 were all identical. The sire of horse 2 was heterozygous for the base change. The sire, dam, and siblings were clinically normal.

Comparisons of the cDNA sequences that encode α_IIb and β₃ in humans and dogs have been reported. ^{9, 10} Horse cDNA sequence encoding α_IIb was found to be more similar to dog (87%) than to human (82%), although when translated to the amino acid level, the percent identity with dog and human were very similar (81 and 82%, respectively). Murine monoclonal antibodies generated against human α_IIb do not cross-react with either the canine or equine platelet integrin subunit (Boudreaux, personal observation). Horse cDNA sequences encoding β₃ were more similar to human and dog sequences at both the nucleotide and amino acid level, with identities ranging from 92 to 95%. In spite of this similarity, monoclonal antibodies to human β₃ that cross-react with the canine subunit do not recognize horse β₃ (Boudreaux, personal observation). The reason for this discrepancy is not known but may be related to the conformation the β₃ subunit achieves when it is complexed with α_IIb on the platelet surface.

GT has been well characterized at the functional, biochemical, and molecular level in humans and dogs. In dogs, two different mutations, both in the gene encoding α_IIb, have been reported. ^{1, 11} According to the human GT database, maintained by the Mount Sinai School of Medicine, over 60 mutations have been documented in the gene encoding α_IIb and over 40 mutations have been documented in the gene encoding β₃ (http://sinaicentral.mssm.edu/intranet/research/glanzmann/menu). Over half of the reported cases of GT in humans involving the gene encoding α_IIb are classified as compound heterozygotes. The findings in this study indicate the likely cause of GT in horse 1 is a single nucleotide change in codon 41 in exon 2 of the gene encoding α_IIb. This area of the gene is highly conserved among humans, dogs, and horses. While cDNA sequence from horse 2 had the identical base change identified in horse 1, genomic DNA sequence was heterozygous for this base change. This suggests that horse 2 is likely a compound heterozygote. The finding that only the sire and not the dam was heterozygous for this base change further confirms that horse 2 is likely a compound heterozygote. The dam is likely heterozygous for the unidentified mutation. The other unidentified mutation apparently results in either total lack of expression or extreme instability of mRNA coding for α_IIb, such that message is not well represented in RT-PCR products of platelet RNA. Thus, this mutation may be located within the promoter region or within an intron or other nontranslated portion of the α_IIb gene. However, a nonsense mutation within the coding portion of the gene could also result in instability of the mRNA, via nonsense-mediated decay. ¹⁸

The base change documented in exon 2 of horses 1 and 2 would be predicted to result in the change of encoded amino acid 41 from an arginine to a proline. The marked difference in structure and charge of proline compared with arginine, as well as the presence of two adjacent prolines at positions 40 and 41, would be predicted to result in marked instability of the beta-propeller region of the encoded protein. A similar missense mutation in exon 2 of the gene encoding α_IIb has been described in humans. ¹⁹ In this case, a single nucleotide change resulted in the substitution of a proline for a leucine at amino acid position 55. Experiments demonstrated that the mutation resulted in severe impairment of expression of the α_IIbβ₃ complex on the surface of transfected COS7 cells. The authors hypothesized that the substituted proline caused an aberrant conformation in the encoded protein that prevented association of α_IIb with β₃, with ultimate lack of expression of the complex on the surface of cells.

This is the first characterization of cDNA sequences encoding platelet α_IIb and β₃ in normal horses and in two horses with GT. Future studies will be aimed at identifying the second mutation in the apparently compound heterozygote horse 2. Horses that have platelet-type hemorrhage, particularly epistaxis, that do not have thrombocytopenia, vasculitis, or von Willebrand disease, are candidates for the diagnosis of GT. Molecular-based screening assays, based on molecular defects identified in horses with GT, will greatly facilitate the identification of GT in horses with otherwise unexplained platelet-type bleeding.