Report

Human Hair Growth Deficiency Is Linked to a Genetic Defect in the Phospholipase Gene LIPH

See allHide authors and affiliations

Science  10 Nov 2006:
Vol. 314, Issue 5801, pp. 982-985
DOI: 10.1126/science.1133276

Abstract

The molecular mechanisms controlling human hair growth and scalp hair loss are poorly understood. By screening about 350,000 individuals in two populations from the Volga-Ural region of Russia, we identified a gene mutation in families who show an inherited form of hair loss and a hair growth defect. Affected individuals were homozygous for a deletion in the LIPH gene on chromosome 3q27, caused by short interspersed nuclear element–retrotransposon–mediated recombination. The LIPH gene is expressed in hair follicles and encodes a phospholipase called lipase H (alternatively known as membrane-associated phosphatidic acid–selective phospholipase A1α), an enzyme that regulates the production of bioactive lipids. These results suggest that lipase H participates in hair growth and development.

Mammalian hair follicles are self-renewing organs that represent interesting models for the regulation of stem cells. Hair follicles cycle through periods of growth (anagen), involution (catagen), and rest (telogen) before regenerating at the onset of a new anagen growth phase (13). Hair follicle stem cells, permanent residents of the stem cell niche called the “bulge,” communicate with the underlying dermal papilla cells and proliferate at anagen onset to generate the progenitor matrix cells required for new hair growth (4). The molecules that control morphogenesis and cycling of hair follicles and the mechanisms underlying hair loss are poorly understood. However, genetic studies of rare familial cases of alopecia (hair loss on the scalp) and hypotrichosis (deficiencies of hair growth) have yielded important information about some of the genes controlling hair growth, including human hairless, desmoglein 4 (DSG4), and corneodesmosin (CDSN) (57).

Previously, we described individuals within the aboriginal Finno-Ugric population of Russia with a genetic form of hair growth deficiency (8, 9). To identify the genetic defect for this condition, we have now studied two ethnic groups of mixed Caucasian and Mongoloid origin living in the Volga-Ural region of Russia (Mari El and Chuvash). The Mari population belongs to the Finno-Ugric linguistic group and the Chuvash population to the Turks linguistic group. The ancestors of the Chuvash population were probably Volga Bulgars, extruded by Mongols from Volga Bulgaria, who settled in the territory occupied by the Mari ancestral populations. We analyzed 50 families with hypotrichosis (14 from Mari and 36 from Chuvash) identified in a genetic epidemiological study of 171,500 Mari individuals and 178,722 Chuvash individuals (see supporting online material).

Affected individuals were characterized by deficiencies of hair growth on the scalp and body starting at birth, but showed no other pathologies. The growth of scalp hairs was retarded or arrested, leading to short hair length. Hair loss on the scalp was occasionally seen in children and progressed with age (Fig. 1A and fig. S1). Histopathological analysis revealed abnormal morphology of hair follicles and dystrophy and fragility of the hairs in analyzed individuals (Fig. 1, B to D).

Fig. 1.

Clinical presentation of hair growth defect and alopecia in Chuvash individuals. (A) Example of hypotrichosis and alopecia in a female adolescent (left) and a male adolescent (right). Notethesparseandshort hairs on the scalp (shown with permission from the subjects). The phenotype is variable in males and females of different ages and can progresses to alopecia in adults. (B) Hair fibers from affected individuals showed common dystrophic structural alterations and signs of fracture and fragility. Scale bars, 100 μm. (C and D) Hair follicle histology in an affected individual (C) and in an unaffected control subject (D). Scale bars, 50 μm. In many follicles, the lower follicular infundibulum above the insertion of the sebaceous gland shows marked dilation with epithelial thinning and abnormal keratinization (brackets). There is a loss of the normal granular layer, with premature differentiation of the epithelium and a remarkable parakeratotic plug.

The parents of affected individuals were normal, and the segregation frequency suggested an autosomal-recessive form of inheritance. We conducted primary genotyping in Mari families with a set of STR (simple tandem repeat) markers selected arbitrarily or from loci implicated in hair defects of humans or rodents. This analysis revealed a potential linkage of the disorder to chromosome 3q26-27. Subsequent homozygous mapping in Mari families with known STR markers in this chromosomal region localized the mutant gene to a ∼2.26-cM interval between D3S1617 and D3S3583 (Fig. 2A and figs. S2 to S4). The mapped interval contained 17 annotated genes, none of which had been previously implicated in hair growth, and some of the genes were excluded by direct mutation analysis of affected individuals. The discovery of shared common STR haplotypes on 3q27 in unrelated affected individuals indicated that the mutation in the two populations likely arose from a common founder. We speculated that an examination of high-density STR maps on mutant chromosomes of individuals that are separated by many generations might critically shorten the genomic interval. We isolated novel STRs and performed genotyping of 17 STR markers overlapping ∼5 cM between D3S1571 and D3S1262. Maximum two-point logarithm of the odds ratio for linkage (lod) scores were found for two markers, (CA5) (Zmax = 11.98, θ = 0.00) and D3S1530 (Zmax = 9.09, θ = 0.00), located near each other. The analysis of shared and recombinant haplotypes in Mari families assigned the mutant gene to the 0.89-cM interval flanked by the D3S3592 and D3S1530 markers (Fig. 2 and fig. S4). By combining these data with the haplotypes in Chuvash families, we localized the gene to a ∼350-kb region encompassing four protein-coding genes (MAP3K13, TMEM41A, LIPH, and SENP2) (Fig. 2).

Fig. 2.

Identification of a mutant gene for hypotrichosis by positional cloning. (A) The STR genotyping identified a linkage to the 3q27 region. The genetic map and genes located in the genomic region are shown. The markers in bold are STR markers newly identified in this study. The recombination events define (CA5) as centromeric and D3S1530 as telomeric boundaries of the minimal genetic interval. The gray and black bars span the haplotypes shared by Mari mutant carriers, and the black bar denotes the region for haplotypes shared by all affected individuals in the Mari and Chuvash populations. Four protein-coding genes and a transcribed sequence (LOC647276) were annotated in this region. Sequence analysis of the LIPH gene revealed that exon 4 and flanking intronic sequences were deleted in affected individuals. The locations of sites for putative catalytic amino acid residues in LIPH exons are shown by arrows. Introns 3 and 5 contain multiple copies of SINEs. Shown is an uneven recombination event between Alu repeat copies leading to the deletion. (B) Alignment of LIPH amino acid sequence with other members of the lipase family. Human paralogous genes are presented in the alignment. Exon 4 encodes a highly conserved domain containing evolutionarily invariant amino acid residues. The Asp178 (exon 4), Ser154 (exon 3), and His248 (exon 6) of LIPH belong to a catalytic triad conserved in lipases (fig. S7). LIPH, NP_640341; LIPI, NP_945347; PS-PLA1, NP_056984; LIPG, NP_006024; LPL, NP_000228; LIPC, NP_000227; PNLIP, NP_000927; PNLIPRP1, NP_006220; PNLIPRP2, NP_005387.

We sequenced the entire coding regions of the MAP3K13, TMEM41A, SENP2, and LIPH genes, including exon-intron junctions, in selected affected individuals and found no disease-associated mutations. We noted, however, that exon 4 of the LIPH gene was not amplified from DNA samples of affected individuals using flanking intronic polymerase chain reaction (PCR) primers. A PCR product derived from LIPH exon 4 was detected in the parents but not in any affected individuals, suggesting that the deletion of exon 4 existed in a heterozygous state in the parents and in a homozygous state in individuals with hypotrichosis.

To map the deletion breakpoints, we designed a series of primers in the genomic region between exon 3 and exon 5, generated PCR products of mutant alleles harboring the deletion, and performed direct sequencing. This mapping of breakpoints determined that a ∼985-bp deletion eliminates exon 4 and flanking intronic sequences. We then examined all families and confirmed this deletion in the homozygous state in affected individuals and in the heterozygous state in their parents (Figs. 2 and 3). The genomic region between exon 3 and exon 5 is unusually densely saturated by SINE (short interspersed nuclear element) repeats dispersed in the human genome (10). Sequencing of PCR products from the deletion and from wild-type alleles predicts that the deletion is a result of unequal recombination between copies of Alu retrotransposons flanking exon 4. The recombination event presumably occurred between highly homologous 5′ regions of the two members of distant subfamilies of Alu elements (Fig. 2A).

Fig. 3.

(A) Examples of LIPH exon 4 deletion identified in affected individuals in Mari (upper panel) and Chuvash (lower panel) populations. PCR fragments in agarose gels are shown. Arrows indicate the upper band representing a normal allele and the lower band of the mutant deletion allele. *, heteroduplex PCR fragment observed in heterozygotes. All affected individuals are homozygous for exon 4 deletion. (B) Reverse transcription PCR (RT-PCR) analysis shows LIPH transcript missing exon 4 in cultured hair follicle bulge keratinocytes from an affected individual (lane 1), LIPH wild-type transcript (lane 2), and negative control (no template) (lane 3). (C) Microphotograph of hair follicle showing sections used for RNA preparation. Scale bar, 100 μm. (D) Semiquantitative RT-PCR analysis of LIPH mRNA in the hair follicles in comparison with other tissues. The KRT15 (keratin 15) gene highly expressed in hair follicles and the GAPDH (glyceraldehyde-3-phosphate dehydrogenase) housekeeping gene were used as controls. Prominent expression of LIPH was observed in a hair follicle bulge in anagen (an. bulge), the active phase of hair growth, but not in dermal papilla (DP) (see also supporting online material and fig. S8).

To determine the frequency of the mutant allele, we tested 2292 chromosomes in population samples collected irrespective of the hypotrichosis phenotype from Volga-Ural and other regions of Russia. As anticipated, we found heterozygous individuals with the LIPH deletion in populations of Chuvash (the mutant allele frequency P = 0.033) and Mari (P = 0.030) origin. No mutant allele was found in other Finno-Ugric populations (Udmurths) or Russian populations from distant geographic regions. The data suggest that there are more than 98,000 heterozygous mutant carriers and 1,500 homozygous affected individuals in populations of Mari and Chuvash descent. The number of mutation carriers in this region may be higher because of the mutant gene flow to other ethnic groups (SOM Text).

Deletion of exon 4 does not alter the reading frame of the LIPH gene; however, the deletion eliminates an evolutionarily conserved domain in the predicted protein (Fig. 2 and figs. S6 and S7). The protein product, lipase H (alternatively called mPA-PLA1α), has a striking sequence similarity to phospholipases and members of the large triglyceride lipase family. Like all lipases, lipase H contains a putative catalytic amino acid triad: Ser154, Asp178, and His248 (11, 12). Replacement of the amino acid triad residues by site-directed mutagenesis abolishes enzymatic activity of lipases (1215). The LIPH mutation associated with hypotrichosis deletes the critical Asp178, as well as several other evolutionarily conserved amino acid residues (Fig. 2B and fig. S7).

It has been postulated that phospholipases generate lysophosphatidic acids (LPA, 1- or 2-acyl-lysophosphatidic acids) from phospholipids (16, 17). LPA has been described as an extracellular mediator of many biological functions, such as proliferation, antiapoptotic activity, and cytoskeletal organization, and appears to signal through at least five members of a family of G protein–coupled receptors (12, 1722). Lipase H is highly homologous to lipase I (LIPI) and phosphatidylserine-specific phospholipase A1 (PS-PLA1). These three proteins also share common structures of short lid domains and partially deleted β9 loop domains that probably determine the specificity of their phospholipase activity in the production of LPA (14, 23) (figs. S7 and S9).

To elucidate whether the LIPH gene is expressed in hair follicle development, we analyzed mRNA transcripts isolated from human hair follicles and other tissues. The expression of LIPH, but not LIPI and PS-PLA1, was prominent in hair follicles, including the stem cell–rich bulge region (Fig. 3 and fig. S8). These data further indicate the importance of LIPH in normal hair formation and growth.

The physiological function of LIPH has not yet been elucidated. We speculate that intragenic deletion of the LIPH (loss-of-function mutation) abolishes the enzymatic activity of lipase H and diminishes the production of LPA mediators in hair follicles. Such lipid mediators may affect the migration, differentiation, or proliferation of keratinocytes, culminating in arrest of hair growth (fig. S9). However, LPA-independent mechanisms and other activities, for example, intracellular functioning of LIPH, cannot be ruled out. Like age-related hair loss in the general population, the hypotrichosis and alopecia described here are not associated with other pathologies, and they progress with age. The identification of a genetic defect in LIPH suggests that this enzyme regulates hair growth and therefore may be a potential target for the development of a therapeutic agent for the control of hair loss or growth.

Supporting Online Material

www.sciencemag.org/cgi/content/full/314/5801/982/DC1

Materials and Methods

SOM Text

Figs. S1 to S10

Table S1

References

References and Notes

Stay Connected to Science

Navigate This Article