To the Editor:

Humans are a diploid species. For any particular region (locus) of the genome, we have two copies (alleles) of DNA information. However, the majority of us carry stretches of DNA for which we cannot delineate the maternal and paternal copies. These are called loss-of-heterozygosity (LOH) events, also known as runs of homozygosity (1). LOH events bring all variants in the locus into a homozygote state and may thus reveal the effects of deleterious recessive mutations. In the laboratory, we can detect LOH by DNA sequencing or genotyping methods. In cancer, LOHs are commonly observed as the result of somatic deletion of one of the alleles (hemizygous deletions). In this situation, there is a loss of allelic dosage where one allele, instead of two alleles, is present. The same laboratory methods can also identify LOHs that are copy neutral, meaning that the two copies are present. This occurs when identical chromosomal segments originating from a common ancestor are inherited from both parents. Copy-neutral LOHs can have serious implications for disease risk. Mating among relatives (consanguinity) is known to increase the number and size of copy-neutral LOHs, as well as the incidence of autosomal recessive disease.

An alpha-1 antitrypsin deficiency (AATD) is caused by mutations in the SERPINA1 (serpin family A member 1) gene on chromosome 14. SERPINA1 encodes the protease inhibitor alpha-1 antitrypsin (AAT). AAT serum concentration is inherited as a codominant trait (i.e., both alleles contribute to the phenotype). Regarding association with lung disease, AATD was historically accepted as a recessive condition, where both alleles had to be altered to cause AAT serum concentrations to drop below a critical threshold needed to protect the lung. However, this view has been challenged by many recent studies confirming increased risk of emphysema and other conditions, such as hepatic diseases, in PiZ heterozygous individuals who smoke (2, 3). Several mutations in SERPINA1 are known to cause AATD (4, 5). The most common mutations are E264V and E342K, which are more commonly known as the S and Z alleles, respectively. As the diagnostic methods for AATD transition to DNA sequencing, an increasing number of null mutations are also detected (6–8). Null mutations are the result of different types of genetic variants (missenses, deletions, insertions, etc.), leading to the absence or near absence of AAT production. To the best of our knowledge, no case of severe AATD has been reported as the result of a single null variant inherited on a copy-neutral LOH event.

As part of our clinical screening for AATD, we encountered an unusual case. The patient was a female with a smoking history of 2 pack-years. Initial bronchodilator reversibility led to the diagnosis of severe asthma that rapidly evolved into a fixed airway obstruction at 33 years of age (forced expiratory volume in 1 s [FEV₁] = 33% predicted; FEV₁/forced vital capacity [FVC]  = 44%; Dl_CO = 65% predicted). Despite optimal asthma treatment, her lung function continued to decline, with an FEV₁ of 25% predicted, Dl_CO of 52% predicted, and chest computed tomography showing panlobular emphysema at the age of 39. On the basis of clinical and imaging features reminiscent of AATD, we decided to investigate using both serum measurement of AAT and SERPINA1 DNA sequencing. The serum AAT concentrations were undetectable by immunoturbidimetry (<0.16 g/L) and at 0.004 g/L by ELISA. DNA analysis revealed a rare null mutation that consists of an insertion of a “T” nucleotide at position 353 of the mature protein, which causes a frameshift and a premature stop codon 24 amino acids downstream (Figure E1 in the data supplement). On the basis of genetic nomenclature, this mutation is thus called Leu353Phe_fsTer24 (rs763023697). Using the historical protease inhibitor system nomenclature, this mutation on the M3 background was first reported in Portugal as the Q0_Ourém (9). We previously reported heterozygote Q0_Ourém patients in our French-Canadian population (10). Accordingly, the presence of this mutation by itself was not unusual, but the odds of being homozygous for this rare mutation are very small. The patient was also homozygous for the E376D mutation (M3 variant on the PI system), which is consistent with our previous study showing that Q0_Ourém is transmitted on the M3 background (10). On the basis of these results, we performed additional sequencing of DNA fragments covering introns 2, 3, and 4 of SERPINA1. We found four additional intronic and in silico–predicted nondeleterious variants. Interestingly, the patient was again homozygous for all these variants. Our initial impression was that only one of the alleles was amplified and sequenced because of a large deletion affecting the entire SERPINA1 gene as observed previously (11).

To investigate this possibility, we used a SNP array; (i.e., the same genotyping technology that is used to perform genomewide association studies) (12). However, instead of looking at all SNPs throughout the human genome, we focused our analyses on the chromosome-14 region where the SERPINA1 gene is located. For a genomic segment of 2 Mb, the heterozygous state completely disappeared (Figure 1; no SNP with a B allele frequency of approximately 0.5). However, there was no change in the log R ratio. Together, this is indicative of a copy-neutral LOH event, meaning that all variants found in this 2-Mb window are homozygous for this patient. Unfortunately, this 2-Mb segment of DNA carries the Leu353Phe_fsTer24 mutation, and the patient thus features severe AATD.

Figure 1. Copy-neutral loss of heterozygosity (LOH) identified in the patient. The upper part shows the single-nucleotide polymorphism (SNP) array results across the region of chromosome 14. Each blue dot represents a SNP. SNPs are plotted on the basis of their position on chromosome 14 on the x-axis against the B allele frequency and the log R ratio (15). Across a window of two megabases (Mb), delimitated by the red rectangle, the heterozygote state has completely collapsed to the homozygote axes, which is indicative of a LOH and depicted by no SNP with a B allele frequency of around 0.5 on the y-axis. In the same window, we observed no change in the log R ratio, depicted by all SNPs having a value of approximately 0 on the y-axis, which is indicative of the presence of two chromosomes. Genes located in this region are illustrated at the bottom and include SERPINA1. Positions are based on GRCh37/hg19 genome assembly.

Download Figure | Download Powerpoint

[More] [Minimize]

Family history data revealed that the patient is the offspring of second cousins (Figure E2). At the genomewide scale, we found 25 additional LOH events in this patient, which strongly excluded the possibility of uniparental disomy (both copies of a segment of chromosome inherited from one parent) at the SERPINA1 locus. The sum of the lengths of all LOH regions is 92.4 Mb, which is consistent with the range expected in the offspring of second cousins (13).

In conclusion, we report the first case of AATD caused by a null variant inherited on a copy-neutral LOH event. The prevalence of LOH events spanning SERPINA1 is currently unknown. It is likely to be rare at this locus in the general population (14), but higher incidences may be observed in areas where consanguinity is more prevalent or in isolated populations. Conventional testing for AATD (AAT serum concentrations and protease inhibitor phenotyping) and genetic testing (targeted genotyping and sequencing) do not detect the presence of LOH. This raises the possibility that LOH events underlie a minority of more common AATD genotypes, such as ZZ homozygotes. It should be emphasized that there are no direct clinical implications of the Q0_Ourém/Q0_Ourém or ZZ genotype with or without LOH as the genetic mechanism underpinning AATD. However, the finding remains important, as it highlights a new genetic mechanism causing AATD.