Nucleotide Polymorphisms in the α₂ Gene Define Multiple Alleles That Are Associated With Differences in Platelet α₂β₁ Density

INTEGRINS ARE heterodimeric molecules composed of noncovalently associated α and β subunits that mediate cell-cell and cell-matrix adhesion.1 The integrin receptor for collagen/laminin, α₂β₁ (also known as the platelet membrane glycoprotein Ia-IIa complex,2 the very late activation antigen-2 [VLA-2],3 and the class II extracellular matrix receptor [ECMII] ⁴), is expressed on a wide variety of cell types, including megakaryocytes, platelets, fibroblasts, endothelial cells, and epithelial cells.5Although α₂β₁ serves as a collagen receptor on platelets and fibroblasts,6 it functions as both a collagen and laminin receptor on endothelial cells and on many epithelial cell types.7 

Previous studies in our laboratory have shown that platelet levels of α₂β₁ vary significantly among normal individuals, whereas the levels of other integrins do not.⁸ Because α₂β₁ mediates platelet adhesion to collagen in vivo, variation in its expression levels could have a significant impact on platelet function. Significantly, our recent analyses have shown that DNA sequence polymorphisms in the α₂ gene are linked to expression levels of α₂β₁ on platelets.9 The DNA sequence variants identified include two conservative changes in the amino acid coding region of α₂ at nucleotides 807 (TTT/TTC at codon Phe²²⁴) and 873 (ACA/ACG at codon Thr²⁴⁶) of the cDNA sequence. Although these particular variants do not change the amino acid sequence of the α₂protein, we have found a significant correlation between these DNA sequence polymorphisms and expression levels of α₂β₁. We have found that the 807C/873G sequences are associated with lower levels of α₂β₁, while the 807T/873A sequences are associated with higher levels of this integrin. In addition, familial studies confirm that regulation of α₂β₁levels is an inherited trait linked to these two silent polymorphisms. These alleles were referred to as 807C (for the 807C/873G pair) and 807T (for the 807T/873A pair).

In this report, we have extended our analysis of the α₂gene to include the introns surrounding the nucleotide polymorphisms at bp 807 and 873. As a result of this work, we are now able to define three α₂ gene alleles; two of these alleles are associated with lower levels of α₂β₁, whereas one is associated with higher levels of this integrin. These allelic differences in α₂β₁ levels could modulate platelet function in vivo by influencing the critical initial phase of stabilized platelet adhesion to collagen, which is mediated by α₂β₁. Indeed, we show here that differences in α₂β₁ receptor levels are reflected in the rate of platelet attachment to type I collagen under conditions of high shear.

The polymorphism at bp 837 was identified after analysis of mRNA or DNA from 85 individuals. The mRNA was amplified and sequenced, whereas the DNA was analyzed by Southern dot blot hybridization for determination of the base at position 837. Procedures were performed as previously described.9 Oligonucleotides used for dot blot analysis of the 837 sequence were: 837 C: GCTGAATAGGCATATTT; 837 T: GCTGAATAAGCATATTT.

DNA was either sequenced manually by the dideoxy termination method using Sequenase 2.0 (US Biochemical Corp, Cleveland, OH) or with an automated sequencer at The Scripps Research Institute DNA Core Laboratory.

The primers used for amplification possessed either an Xba I site or an Xho I site to facilitate subcloning into pGEM (Promega, Madison, WI). Intron G was amplified and isolated as previously described.9 Intron F was amplified using the following primer pair: 5′ primer (cDNA bp 634-665): GAACTCGAGGTACAAGGCCTTGATATAGGCCC; 3′ primer (cDNA bp 764-786): AGGTCTAGACCATATTGGGATGTCTGGGATG. Intron H was amplified using the following primer pair: 5′ primer (intron G bp 3506-3532): AATCTCGAGCGAATACTGGGATAAATACATGCAC; 3′ primer (cDNA bp 1090-1114): TCCTCTAGACCCAGCCTTTTCTAGTAGAGCTGC.

The polymerase chain reaction (PCR) products were ligated into pGEM (Promega) after digestion with Xba I and Xho I. Intron F was found to possess an internal Xba I site. Therefore, each segment of intron F was subcloned into pGEM separately.

Determination of Br genotype was performed as described.10Sixty-three individuals were typed for this polymorphism; of these, 13 were heterozygous for this polymorphism, whereas the remainder were homozygous for the Br^b polymorphism.

The ∼600-bp segment of intron G encompassing the Bgl II andNde I sites was amplified from genomic DNA using the following primer pair: 5′ primer (intron G bp 2789-2812): GATTTAACTTTCCCGACTGCCTTC; 3′ primer (intron G bp 3346-3369): CATAGGTTTTTGGGGAACAGGTGG.

The PCR product (5 μL) was digested in a 15 μL reaction volume using 0.5 μL of Bgl II or Nde I (NEB) in the recommended reaction buffer at 37°C overnight. Reaction products were analyzed on a 1.4% agarose gel.

Acid soluble type I collagen from human placenta (Sigma, St Louis, MO) was diluted to a concentration of 200 mg/mL in phosphate-buffered saline (PBS), pH 7.4, and applied evenly over a horizontal glass cover slip (Corning, Inc, Corning, NY; 24 mm × 50 mm), covering all but the first 10 mm, which remained uncoated to facilitate handling. Coated cover slips were then placed in a humid environment at room temperature for 60 minutes. Excess collagen was removed by four sequential rinses with PBS, pH 7.4, and assembled in the flow chamber. Blocking the cover slips with bovine serum albumin (0.1 mg/mL) did not affect initial platelet adhesion or subsequent thrombus formation, and uncoated cover slips did not support platelet adhesion.11,12 

Platelet interaction with immobilized collagen was studied using a modification of a Hele-Shaw flow chamber described elsewhere.11,12 Collagen-coated cover slips formed the lower surface of the chamber with a flow path height of 254 mm (determined by a silicone rubber gasket). The flow chamber was assembled and filled with PBS, pH 7.4. A syringe pump (Harvard Apparatus Inc, Holliston, MA) was used to aspirate blood through the flow chamber. A flow rate of 1.94 mL/min produced a wall shear rate of 1,500/s at the inlet of the flow chamber. All measurements of platelet adhesion and thrombus formation were made at a position adjacent to the inlet of the chamber so as to avoid prior exposure of flowing platelets to the thrombogenic surface and preformed thrombi. Platelets were labeled in whole blood by direct incubation with the fluorescent dye mepacrine (quinacrine dihydrochloride; 10 mmol/L, final concentration). Although this dye also labels leukocytes, these cells could be readily distinguished from platelets by their relatively large size and scarcity; moreover, permanent leukocyte attachment to collagen was negligible at wall shear rates above 500/s. Red cells were not visualized due to fluorescence quenching by hemoglobin. Mepacrine concentrates in the dense granules of platelets and has no effect on normal platelet function at the concentration used.13 Platelet secretion after adhesion does not prevent their visualization. Furthermore previous studies have shown that mepacrine does not interfere with platelet adhesion.12 The flow chamber, mounted on an epifluorescence microscope (Axiovert 135M inverted microscope, Carl Zeiss Inc, New York, NY), allowed direct visualization in real time of the platelet adhesion process, which was recorded on a video cassette recorder.

The total area occupied by adherent platelets in an area of 16,384 mm² was measured by capturing images from the video tape using a frame grabber (Matrox Image LC; Matrox Electronics System Ltd, Dorval, Quebec, Canada) and processing using the Metamorph software package (Universal Imaging Corporation, Westchester, PA). A threshold was applied to the image to distinguish platelets from the background. The microscope settings (including contrast, brightness, and magnification settings) were maintained at constant values to facilitate valid comparisons between different experiments.

We have previously described the variation in α₂β₁ levels that exists among individuals. In studying this variability in expression levels, our analysis of mRNA from six individuals expressing different levels of α₂β₁ revealed two linked nucleotide polymorphisms, separated by almost seventy nucleotides, at bp 807 and 873 in the α₂ coding region. These were the only two nucleotide polymorphisms identified within the ∼3.5-kb α₂ coding region that consistently varied among the samples studied. Based on this work, subsequent analyses of 30 individuals showed that the 807/873 nucleotide polymorphisms were associated with differences in expression levels of α₂β₁; the 807C/873G pair was found to be associated with lower levels of α₂β₁ , the 807T/873A pair with higher levels of this integrin.

In the present study, we have continued our analysis of the α₂ gene to determine if additional sequence polymorphisms exist that are associated with the 807 and 873 polymorphisms and with expression levels of α₂β₁. Analysis of a larger group of individuals has now led to the identification of an additional, rarer nucleotide polymorphism (C or T) in the coding region of the α₂ gene, at bp 837. As with the polymorphisms at bp 807 and 873, the polymorphism at bp 837 does not change the α₂ coding sequence. In this variant, a C rather than a T is found at bp 837 in a small subset of individuals (Fig 1). In our study group, 13 out of 85 individuals screened were heterozygous C/T at this position; all 13 of these individuals were either heterozygous or homozygous for 807C/873G, indicating strong linkage with the 837C polymorphism.

This nucleotide variation therefore defines a third allele of the α₂ gene, which is present at a frequency of approximately 8% in the population. It is striking that the three nucleotide polymorphisms we have identified are clustered in one segment of the α₂ coding region. For this reason, we extended our analysis of the α₂ gene to include the three introns flanking bp 807, 837, and 873 (introns F, G, and H). We have cloned these introns, and the sequences of this region have been deposited with GenBank (Accession No. AF035968).

We have identified several sequence variations in the ∼4 kb intron that separates bp 807 and 873 (intron G). Each of these additional nucleotide polymorphisms is always found associated with the particular α₂ allele indicated in Fig 1, thereby establishing their linkage to the 807 and 873 polymorphisms. Of particular interest for screening purposes (see below) is a Bgl II restriction site created by the nucleotide sequence present in intron G of allele 1. As depicted in Table 1, this Bgl II site is found associated only with allele 1. Seventy nucleotides upstream of the Bgl II site, we have also identified anNde I site, present in alleles 1 and 2, but are absent in allele 3. Interestingly, this nucleotide variation is linked to the polymorphism at bp 837, where only individuals carrying a C at bp 837 were found to lack the Nde I site. The previously defined Br polymorphism at nucleotide 1648¹⁰ also appears to be linked to these polymorphisms; only individuals carrying a C at bp 837 were found to carry the Br^a polymorphism.

The linkage associations between the polymorphisms at bp 807/873/837, the Nde I site, and the Br polymorphisms are given in Table 2. Two additional nucleotide polymorphisms linked to bp 807/873 were also identified, one near the 5′ end of intron G and one in the 300 bp intron that lies downstream of bp 873 (intron H), although our analysis is limited to only a few individuals at the present time. Finally, our preliminary data indicate that nine additional base variations in intron F, which lies upstream of bp 807, may be linked to the already defined alleles. However, these particular base changes (determined for just one 807T allele and one 807C allele) need to be confirmed among a larger pool of 807C and 807T individuals.

Thus, three α₂ gene alleles, defined by eight nucleotide polymorphisms, have now been defined. We had previously shown that alleles 1 and 2 of the α₂ gene, defined only by the polymorphisms at bp 807 and 873, were associated with high or low α₂β₁ levels, respectively. The third allele we now define carries the 807C/873G polymorphisms and would have therefore been associated with lower levels of α₂β₁ in our previous studies. Individuals homozygous for allele 3 have not yet been identified by us, precluding a direct analysis of α₂β₁ levels associated with that genotype. Nonetheless, the clustering of nucleotide polymorphisms identified so far suggests that this region of the α₂ gene might harbor elements important in the regulation of α₂ expression.

We have developed a strategy to type individuals for each of the three alleles using the polymorphic Bgl II and Nde I restriction sites described above. The region of intron G encompassing these sites was amplified from genomic DNA as described in Materials and Methods, and the resulting PCR products were digested with Bgl II orNde I. Analysis by agarose gel electrophoresis resulted in clearly resolved patterns from which the genotype of individuals could be readily determined (Fig 2). As seen in panel A, the 600-bp PCR product amplified from individuals expressing alleles 2 and/or 3 (lanes 5 and 6) or from the control sequence (lane 7) remained undigested by Bgl II. The PCR product produced from individuals homozygous for allele 1 (lanes 1 and 2) or from the control sequence (lane 8) was completely digested byBgl II to produce fragments of 200 bp and 400 bp, whereas that produced from individuals expressing allele 1 and either allele 2 or 3 yielded fragments of 200 bp, 400 bp, and 600 bp (lanes 3 and 4).

The same PCR products were digested with Nde I to permit determination of allele 3. As seen in panel B, only those products generated from individuals carrying allele 3 or from the control sequence remained undigested by Nde I (lanes 3, 7, and 10). As none of the individuals screened were homozygous for allele 3, NdeI in each case cuts some product (excluding the control sequence in lane 7). Thus, we describe here a novel RFLP method for rapidly determining the α₂ genotype at positions 807, 837, and 873 from genomic DNA. This technique therefore provides a simple method for discriminating among the three α₂ gene alleles described above.

Sequence analysis of intron G has further revealed the presence of a unique inverted repeat in this region (Fig 1). Separated by ∼2.5 kb, the inverted repeat sequences, which do not have significant homology with sequences currently in the database, are 275 bp long, have the potential to form a stem loop structure based on sequence analysis, and are present in all of the gene alleles described above.

We compared the behavior of platelets from individuals homozygous for allele 1 with those expressing only alleles 2 or 3 to assess the influence of α₂β₁ receptor density on adhesion to type I collagen. Whole blood was perfused over immobilized type I collagen under high shear rate conditions (1,500/s). The results from one representative comparison between an individual homozygous for allele 1 (807T) and an individual homozygous for allele 2 (807C) are depicted in Fig 3. Coverage of the collagen-coated surface increased over time with platelets from both individuals, and by 3 minutes there was essentially no difference in the total surface coverage between the two samples. However, the rate at which platelet deposition occurred differed significantly between paired donor samples. The results from a comparison of three 807T and three 807C donors are summarized in Fig 4. In the case of the 807T donors, the rate of platelet attachment was significantly faster, particularly within the first 2 minutes of adhesion (Fig 4). Thus, at high shear rates in whole blood, the rate of platelet attachment to type I collagen increases with increasing density of α₂β₁.

We have previously defined two linked, silent polymorphisms in the α₂ gene that are associated with variable expression levels of α₂β₁ on platelets and that define two alleles of the α₂ gene.9 In this report, we describe the identification of five additional nucleotide polymorphisms in the region surrounding bp 807 and 873. These polymorphisms are linked to those previously described and to expression levels of α₂. Indeed, three α₂gene alleles, defined by eight nucleotide polymorphisms, have now been confirmed.

It is not clear whether these sequence changes are themselves involved in differential α₂ expression or if they are merely physically linked to the region responsible for modulating α₂ levels. Studies are currently underway to investigate this question. In the discussion that follows, we describe several possible mechanisms through which these exonic and/or intronic polymorphisms might be linked to expression levels of the α₂β₁ integrin; other mechanisms are certainly possible.

There are a number of ways in which DNA sequence alterations could influence gene expression. For example, changes in specific promoter elements could influence promoter activity, resulting in increased or decreased mRNA levels. Expression of α₂β₁in hematopoietic cells is restricted to megakaryocytes and platelets.14 Zutter et al15 have shown that induced expression of α₂β₁ during megakaryocytic differentiation is due to transcriptional activation of the α₂ gene; no changes in the level of β₁mRNA expression were detected during megakaryocytic differentiation. Approximately 5 kb of the 5′ sequence flanking the α₂ transcriptional start site has been cloned, and regions critical for α₂ expression have been identified.16 Although a strong core promoter region located in the first ∼100 bp upstream of the transcriptional start site was found to be necessary for gene activity, it was not found to be cell-type specific. A silencer element located upstream of this region (between −92 and −351) was found to function in cells of hematopoietic lineage, and enhancer elements identified further upstream (between −1426 and −2592) were found to be megakaryocyte-specific and were required for high-level expression of the α₂ gene in megakaryocytic cells. In addition, a number of sites for transcription factors and other regulatory molecules have been identified in the region upstream of the transcriptional start site. Sequence variations in these or other regulatory elements could have a profound impact on expression levels of α₂.

In addition to promoter variations influencing gene expression, the 3′ untranslated region (3′UTR) of α₂, which is approximately 5 kb, might modulate protein expression; 3′untranslated regions have been shown to influence mRNA stability as well as message translation.17 It is also possible that the polymorphisms at bp 807 and 873 are directly related to allelic differences in α₂ expression. Modulation of expression levels by sequences within the coding region of RNA has been described previously.18-20 Differences in intragenic sequences could also influence α₂ gene expression. There are numerous reports of enhancer or repressor elements located in intronic sequences.21-23 Indeed, the fact that the polymorphisms so far identified are clustered in one region of the α₂ gene suggests that this region might harbor critical control elements involved in the regulation of α₂expression.

In whole blood, under a broad range of shear rates (50 to 1,500/s), α₂β₁ mediates the initial phase of stabilized platelet adhesion to collagen that will lead to accelerated prothrombin conversion on the platelet surface and thrombus formation supported by the integrin α_IIbβ_3.^24-26 However, even when the function of α_IIbβ₃ is completely inhibited and thrombus formation is blocked, α₂β₁-mediated adhesion itself still induces the cellular changes that catalyze prothrombin conversion to the same extent as one would otherwise see in the absence of inhibitors.26 Consequently, α₂β₁ has the potential to play a major role in platelet function in vivo both as a primary mediator of adhesion to collagen and as a primary stimulus for prothrombin conversion at the platelet surface.

As shown here, the density of platelet α₂β₁has an important impact on platelet adhesion to collagen in whole blood even under conditions of high shear rate (1,500/s), such that the rate of adhesion to type I collagen increases with increasing receptor density. This stage of adhesion, occurring between 30 seconds and 3 minutes, is likely to be a critical period of thrombus formation in vivo, because there are many compensatory antithrombotic mechanisms that would come into play to inhibit thrombus expansion or to dissociate the nascent thrombus. Within 3 to 5 minutes, a sufficient number of stimuli have been received, regardless of α₂β₁ density, and mature thrombi begin to cover the collagen surface. The expansion of the initial platelet clusters to form larger thrombi is certainly mediated by the binding of integrin α_IIbβ₃, because macroaggregate formation (but not surface attachment) is inhibited by monoclonal antibodies specific for that integrin.

The expression of α₂β₁ by platelets is critical in promoting platelet adhesion to the subendothelium; adhesion of platelets to collagen is critical for normal platelet activity, in hemostasis, and in wound repair. Hereditary variation in platelet levels of α₂β₁, defined by the existence of multiple alleles of the α₂ gene that are associated with variable α₂β₁ expression levels, could therefore have a significant impact on platelet function, contributing to an increased risk of thrombosis or bleeding in relevant disease states.