Vitiligo is a complex disease in which autoimmune destruction of epidermal melanocytes results inpatches of depigmented white skin. Vitiligo has an estimated prevalence of about 0.2e2% in different populations and approximately 0.4% in the European derived white (EUR) population. The fraction of disease risk attributable to genetic variation, termed heritability, is high, with estimates from family studies in EUR of 0.75e0.83 and from SNP based studies estimated at 0.78. About 70% of genetic risk comes from common genetic variants and about 30% from rare genetic variants. Through candidate gene, genomewide linkage, and genomewide association studies, over 50 vitiligo susceptibility loci have been discovered. These have been combined into a vitiligo polygenic risk score, which has allowed various aspects of vitiligo genetic architecture in the EUR population to be better understood. Vitiligo has thus proved to be a particularly tractable model for investigation
of complex disease genetic architecture. Here, we summarize progress to date including dissection of heritability, discovery of vitiligo susceptibility loci through candidate gene, genomewide linkage, and genomewide association studies, relationships to other autoimmune diseases, polygenic architecture of vitiligo risk, vitiligo triggering, and disease onset, and provide suggestions for future directions.
INTRODUCTION
Vitiligo is a common disease in which autoimmune destruction of epidermal melanocytes results in patches of depigmented white skin. Pigment loss is generally progressive, and until very recently, there were no good treatments. Vitiligo is a complex disease, involving both genetic and environmental components. From the genetic standpoint, vitiligo behaves as a typical polygenic condition, each specific genetic factor making a relatively small individual contribution. Nevertheless, vitiligo polygenicity is relatively low and heritability relatively high compared with those of most other complex traits. As a result, the genetic architecture of vitiligo has been easier to discover and understand than that of most other complex traits (Roberts et al., 2020b). Thus, vitiligo provides a particularly tractable model for investigation of complex disease genetic architecture and provides important lessons for predictive, personalized medicine of complex diseases.
Historical Context
Because of its visually dramatic phenotype, vitiligo was recognized relatively early, probably thousands of years ago (Krauss, 2018), with the first known formal medical description in 1765 (Le Cat, 1765). The fundamental pathological defect in vitiligo was defined about a hundred years later when Kaposi (1879) observed a lack of pigmented cells in the involved lesions. Another key observation came with Addison’s (1855) initial description of Addison’s disease, 2 of his 12 cases additionally having vitiligo as well as pernicious (Addisonian) anemia (Figure 1). DeMowbray (1965) suggested that such co-occurring diseases were autoimmune and, furthermore, involved a shared genetic predisposition. The first specific consideration of vitiligo genetics came in 1950 when Stüttgen (1950) and Teindel (1950) simultaneously described multiplex families having multiple relatives affected with vitiligo. Indeed, Stüttgen’s (1950) suggestion that co-occurrence of vitiligo and autoimmune thyroid disease in his family might reflect simultaneous segregation of both dominant and recessive factors constitutes one of the earliest specific hypotheses of complex inheritance.
In the 1960s, efforts began attempting to identify genes that underlie genetic susceptibility to vitiligo (Roberts and Spritz, 2018, Spritz and Andersen, 2017). The earliest studies analyzed polymorphic blood proteins, followed subsequently by candidate gene association studies, targeted genetic linkage analyses, genome-wide linkage analyses, and finally GWAS. The GWAS approach has been by far the most successful, thus far defining 50 confirmed loci that are associated with vitiligo susceptibility in the European-derived white (EUR) population and several others in other populations, especially in Han Chinese (CHN). For many of these loci, the corresponding genes have now been identified and, in many cases, also the specific causal DNA sequence variants. This has led to a deep understanding of vitiligo pathobiological pathways and relationship to environmental triggers and has suggested potential new approaches to vitiligo treatment.
Vitiligo Epidemiology and Heritability
Vitiligo has an estimated prevalence of about 0.2–2% in different populations (Zhang et al., 2016), approximately 0.4% in the EUR population. Although about 91% of cases are sporadic (Roberts et al., 2019), the frequency of vitiligo is considerably elevated in probands’ close relatives, about 7% in first-degree relatives (Alkhateeb et al., 2003). Overall, about 8% of patients report at least one affected relative, with a non-Mendelian pattern of recurrence characteristic of complex, polygenic inheritance (Laberge et al., 2005).
The fraction of disease risk attributable to genetic variation is termed heritability (h2), with the remainder attributable to the environment. Heritability sets a ceiling on the importance of genes in disease causation and provides a key benchmark of progress toward the completeness of disease gene discovery. Classical approaches to estimating heritability analyze disease recurrence rates in multiplex families or twins (Tenesa and Haley, 2013); here, we term such family-based heritability h2FAM. The recent availability of very large genetic reference panels from multiple human populations (McCarthy et al., 2016), coupled with techniques for deep imputation of millions of untyped common and rare variants (Das et al., 2018), has enabled the estimation of overall genomic similarity among unrelated, singleton (simplex) cases versus controls (Yang et al., 2011, Yang et al., 2015); here, we term such DNA-based estimates of heritability h2SNP.
For vitiligo, estimates of h2FAM and h2SNP are remarkably congruent. Estimates of vitiligo h2FAM from small studies range from about 0.50 to 0.80 in different populations (Arcos-Burgos et al., 2002, Das et al., 1985, Hafez et al., 1983, Zhang et al., 2004) and from a large-scale study in the EUR population was recently estimated as 0.75–0.83, depending on the specific types of relatives studied (Roberts et al., 2020a). These estimates of vitiligo h2FAM are all much higher than those of many other complex diseases, for which h2FAM estimates are typically in the range of 0.3–0.5 (Polderman et al., 2015).
In parallel, vitiligo h2SNP was recently estimated as 0.78 in the EUR population (Roberts et al., 2020a), virtually identical to the estimates of h2FAM in this population. Thus, for vitiligo, essentially all heritability estimated by classical family-based methods can be captured by array-based genotyping and deep imputation of common and rare genomic variation; there is no missing heritability (Manolio et al., 2009). With the progress of imputation and sequencing methods that account for both common genetic variants (risk allele frequency ≥ 0.01), such as those detected by typical GWAS, and rarer genetic variants (risk allele frequency < 0.01), this is also proving to be true for other complex traits (Yang et al., 2015; Wainschtein et al., 20191).
Furthermore, the vitiligo h2SNP estimate can be partitioned into the fraction attributable to common genetic variants versus rarer genetic variants (Roberts et al., 2020a). As shown in Figure 2, in the EUR population, common genetic variants account for about 70% of vitiligo h2SNP, whereas rarer genetic variants in the aggregate account for about 30%. Specific identification of such rare vitiligo susceptibility alleles will likely require family studies or GWAS with very large sample sizes that may be difficult to attain for vitiligo.
Figure 2. Partitioning of vitiligo risk. Heritability studies (Roberts et al., 2020a) enable total vitiligo risk to be partitioned into an environmental component (∼20%) and a genetic component (∼80%). Moreover, total vitiligo genetic risk can be further partitioned in a ∼70% fraction attributable to common genetic variants (MAF ≥ 0.01), such as those detected by GWAS, and a ∼30% fraction that represents rare genetic variants (MAF < 0.01). MAF, minor allele frequency.
Together, these findings show that vitiligo heritability is quite high; about 80% of vitiligo risk is genetically based, with the remaining 20% attributable to environmental factors (Figure 2). Furthermore, of the genetic risk component, about 70% (56% of total risk) comes from common genetic variants and about 30% (24% of total risk) comes from rare variants, with no remaining missing heritability. Similar patterns have recently been observed for other complex human traits (Hartman et al., 20192).
Identification of Vitiligo Susceptibility Genes
Candidate gene studies
The first attempts to identify specific genes that underlie vitiligo susceptibility were candidate gene association studies. Unfortunately, this study design has proven highly prone to false-positive results because of population stratification artifacts, incomplete correction for multiple testing, and publication bias of positive results (Hirschhorn et al., 2002). Accordingly, candidate gene studies are no longer generally accepted as appropriate for de novo identification of susceptibility genes for complex traits. Nevertheless, this approach successfully provided the first indications of involvement in vitiligo susceptibility of two immune-related candidate genes, CTLA4 (Blomhoff et al., 2005, Kemp et al., 1999) and PTPN22 (Cantón et al., 2005), both subsequently confirmed by GWAS results.
Genome-wide linkage studies
Genome-wide mapping methods, including both genome-wide linkage studies and GWAS, are not subject to the problems inherent in candidate gene studies. Genome-wide linkage studies search for genomic regions that cosegregate along with the occurrence of disease among affected relatives in multiplex families. Genetic linkage studies are difficult to accomplish, have relatively modest statistical power and low genetic resolution, and are best suited to detecting uncommon causal variants that exert relatively large effects on disease risk. A total of seven putative vitiligo susceptibility loci were detected by genome-wide linkage analysis (Table 1). Of these, five have been associated with a proposed underlying causal gene: FOXD3 (Alkhateeb et al., 2005), NLRP1 (Jin et al., 2007), PDGFRA (Xu et al., 2010), HLA (Yang et al., 2018), and XBP1 (Ren et al., 2009).
Table 1. Vitiligo Susceptibility Loci Detected by Genome-wide Linkage Studies
| Chromosome | LOD Score | Proposed Gene | Population | References |
|---|---|---|---|---|
| 1p31.3-p32.2 | 5.59 | FOXD3 | EUR | Alkhateeb et al., 2005, Spritz et al., 2004 |
| 4q12-q21 | 4.01 | PDGFRA | CHN | Chen et al., 2005, Xu et al., 2010 |
| 6p21-p22 | 3.73 | HLA | CHN | Liang et al., 2007, Yang et al., 2018 |
| 7q21.11 | 3.73 | Unknown | EUR | Spritz et al., (2004) |
| 8p12 | 3.36 | Unknown | EUR | Spritz et al., (2004) |
| 17p13.3 | 3.07 | NLRP1 | EUR | Jin et al., 2007, Nath et al., 2001, Spritz et al., 2004 |
| 22q12 | 3.26 | XBP1 | CHN | Liang et al., 2007, Ren et al., 2009 |
Abbreviations: EUR, European-derived white population; CHN, Han Chinese population; LOD, logarithm of the odds.
GWAS
GWASs typically search for differential genetic associations in singleton cases versus unrelated controls, interrogating hundreds of thousands or millions of DNA polymorphisms across the genome (Altshuler et al., 2008). Furthermore, the genotypes of millions of additional polymorphisms can be imputed using various genome-wide reference panels (Yang et al., 2011, Yang et al., 2015). Unlike candidate gene association studies, GWAS can adequately control for population stratification artifacts, apply appropriate multiple-testing correction, and further require independent replication; thus, GWAS results have been largely reproducible across multiple studies. Over the past decade, GWASs have become the gold standard for the primary discovery of genes involved in complex traits, providing deep insights into the corresponding underlying biology (Visscher et al., 2017). At least five GWASs have been reported for vitiligo; three in the EUR population (Jin et al., 2016, Jin et al., 2012, Jin et al., 2010a), one in CHN (Quan et al., 2010, Tang et al., 2013), a small GWAS in Japanese (Jin et al., 2015), as well as a small partial GWAS in subjects from the Indian subcontinent (Birlea et al., 2013) based on the Immunochip (Cortes and Brown, 2011). For reasons that are unclear, the greatest yield has come from the GWAS of vitiligo in EUR, in which 50 loci have thus far been identified that contribute to vitiligo risk (Table 2). Several additional loci have been discovered in CHN (Table 2). A number of vitiligo susceptibility loci appear to be shared across populations, although others may not be, suggesting that the general pathobiology of vitiligo is likely similar across different populations and efficacy of vitiligo treatments might thus transcend ethnic populations.
Table 2. Vitiligo Susceptibility Loci Detected by GWAS
| Chromosome | Proposed Gene | Population | OR | References |
|---|---|---|---|---|
| 1p36.23 | RERE | EUR | 1.21 | Jin et al., (2010a) |
| 1p13.2 | PTPN22 | EUR, IND, AR | 1.38 | Jin et al., (2010a) |
| 1q24.3 | FASLG | EUR, CHN | 1.32 | Jin et al., (2016) |
| 1q31.3-q32.1 | PTPRC | EUR | 0.83 | Jin et al., (2016) |
| 2p16.1 | PPP4R3B | EUR | 1.51 | Jin et al., (2016) |
| 2q13 | BCL2L11 | EUR | 1.15 | Jin et al., (2016) |
| 2q24.2 | IFIH1 | EUR | 0.77 | Jin et al., (2012) |
| 2q33.2 | CTLA4 | EUR | 1.18 | Jin et al., (2016) |
| 2q37.3 | FARP2 – STK25 | EUR | 0.80 | Jin et al., (2016) |
| 3p24.3 | UBE2E2 | EUR | 0.87 | Jin et al., (2016) |
| 3p13 | FOXP1 | EUR | 0.80 | Jin et al., (2010b) |
| 3q13.33 | CD80 | EUR | 1.31 | Jin et al., (2012) |
| 3q28 | LPP | EUR | 1.32 | Jin et al., (2010a) |
| 3q29 | FBXO45 – NRROS | EUR, CHN | 0.87 | Jin et al., (2016) |
| 4q24 | PPP3CA | EUR | 0.87 | Jin et al., (2016) |
| 6p25.3 | IRF4 | EUR | 0.75 | Jin et al., (2016) |
| 6p25.2 | SERPINB9 | EUR | 0.79 | Jin et al., (2016) |
| 6p22.1 | MHC class I | |||
| HLA-A1 | EUR, JPN, CHN | 1.53 | Hayashi et al., 2016, Jin et al., 2015, Jin et al., 2010a | |
| HLA-A/HLA- B/HLA-C | CHN | 1.90 | Quan et al., (2010) | |
| MHC class II | ||||
| HLA-DRB1- HLA-DQA12 | EUR | 1.77 | Cavalli et al., 2016, Jin et al., 2016, Jin et al., 2010a | |
| HLA-DQA1 – HLA–DQB13 | EUR | 8.1 | Jin et al., (2019) | |
| HLA-DQB14 | CHN | 1.79 | Yang et al., (2018) | |
| BTNL2 – HLA-DRA | IND | 1.67 | Birlea et al., (2013) | |
| 6q15 | BACH2 | EUR | 1.27 | Jin et al., (2012) |
| 6q27 | RNASET2 – FGFR1OP – CCR6 | EUR, CHN | 0.79 | Jin et al., 2010b, Quan et al., 2010 |
| 7p14.3 | CPVL | EUR | 1.84 | Jin et al., (2010b) |
| 8q24.21 | PVT1 | EUR | 1.17 | Ben et al., (2018) |
| 8q24.22 | SLA | EUR | 1.19 | Jin et al., (2012) |
| 9q33.3 | NEK6 | EUR | 1.15 | Jin et al., (2016) |
| 10p15.1 | IL2RA | EUR | 0.77 | Jin et al., (2010a) |
| 10q21.2 | ARID5B | EUR | 1.18 | Jin et al., (2016) |
| 10q22.1 | SLC29A3 – CDH23 | CHN | 0.88 | Tang et al., (2013) |
| 10q22.3 | ZMIZ1 | CHN | 1.18 | Quan et al., 2010, Sun et al., 2014 |
| 10q25.3 | CASP7 | EUR | 0.82 | Jin et al., (2012) |
| 11p13 | CD44 | EUR | 1.23 | Jin et al., (2012) |
| 11q13.1 | PPP1R14B – PLCB3 – BAD – GPR137 –KCNK4 – TEX40 – ESRRA – TRMT112 – PRDX5 | EUR | 0.87 | Jin et al., (2016) |
| 11q14.3 | TYR | EUR | 0.67 | Jin et al., (2010a) |
| 11q21 | Gene desert | EUR | 1.34 | Jin et al., (2010a) |
| 11q23.3 | DDX6-CXCR5 | CHN | 1.22 | Tang et al., (2013) |
| 12q13.2 | IKZF45 | EUR, CHN | 1.33 | Jin et al., 2010b, Tang et al., 2013 |
| 12q24.12 | SH2B3 | EUR | 0.79 | Jin et al., (2012) |
| 13q14.11 | TNFSF11 | EUR | 1.17 | Jin et al., (2016) |
| 14q12 | GZMB | EUR, CHN | 1.25 | Jin et al., (2010a) |
| 15q12-q13.1 | OCA2 – HERC2 | EUR | 1.37 | Jin et al., (2012) |
| 16q24.3 | MC1R | EUR | 0.71 | Jin et al., (2012) |
| 17q21.2 | KAT2A-HSPB9-RAB5C | EUR | 1.21 | Jin et al., (2016) |
| 18q21.33 | TNFRSF11A | EUR | 1.21 | Jin et al., (2016) |
| 19p13.3 | TICAM1 | EUR | 1.21 | Jin et al., (2016) |
| 19q13.33 | SCAF1-IRF3-BCL2L12 | EUR | 0.84 | Jin et al., (2016) |
| 20q11.22 | RALY – ASIP | EUR | 0.61 | Jin et al., (2016) |
| 20q13.13 | PTPN1 | EUR | 1.15 | Jin et al., (2016) |
| 21q22.3 | UBASH3A | EUR | 1.35 | Jin et al., (2010a) |
| 22q12.3 | C1QTNF6 | EUR | 1.32 | Jin et al., (2010a) |
| 22q13.2 | ZC3H7B – TEF | EUR | 0.79 | Jin et al., (2012) |
| Xp21.3-p21.2 | IL1RAPL1 | EUR | 1.77 | Jin et al., (2016) |
| Xp11.23 | CCDC22-FOXP3-GAGE | EUR, CHN, IND | 0.86 | Jin et al., (2016) |
Abbreviations: AR, Arab; CHN, Han Chinese; EUR, European-derived white; IND, Indian subcontinent; JPN, Japanese population; MHC, major histocompatibility complex.
For associations observed in multiple populations, the first-listed was determined by GWAS and the others by non-GWAS methods.
- 1
-
All three populations have major vitiligo association with HLA-A∗02:01.
- 2
-
Intergenic SNP rs9271597-rs9271600-rs9271601 haplotype associated with vitiligo susceptibility.
- 3
-
Intergenic SNP rs145954018 associated with early-onset vitiligo.
- 4
-
Principal association is with HLA-DQB1∗02:02.
- 5
-
Tang et al. (2013) interpreted this association as possibly representing PMEL.
As shown in Figure 3, the majority of the loci that have been associated with vitiligo encode genes involved in immunoregulation, apoptosis, and melanocyte biology. For a number of these genes, the causal genetic variants have been discovered, providing important insights into biological mechanisms that underlie vitiligo risk attributable to these genes, Furthermore, together, the corresponding proteins define a functional network that suggests a general pathobiological pathway of melanocyte damage, antigen processing and presentation, immune cell activation, and melanocyte targeting and apoptosis (Spritz and Andersen, 2017).
Figure 3. Functional interaction network of proteins encoded by candidate genes associated with vitiligo susceptibility. Unsupervised functional interaction network analysis was performed using STRING, version 11.0 (Szklarczyk et al., 2019) with default settings. Primary nodes were proteins encoded by all the confirmed vitiligo susceptibility loci (Tables 1 and 2). Nodes that shared no edges with any other nodes were then excluded. Edge colors are from STRING: teal, interactions from curated databases; purple, experimentally determined interactions; dark green, gene neighborhood; red, gene fusions; dark blue, gene co-occurrence; light green, text mining; black, coexpression; lavender, protein homology.
Relationship to Other Autoimmune Diseases
Epidemiological studies have shown that the frequencies of several other autoimmune diseases are elevated in patients with vitiligo. In the EUR population, these principally include autoimmune thyroid disease (principally Hashimoto thyroiditis), rheumatoid arthritis, adult-onset type 1 diabetes, pernicious anemia, Addison’s disease, and systemic lupus erythematosus (Alkhateeb et al., 2003, Cunliffe et al., 1968, Laberge et al., 2005). Epidemiological studies of autoimmune diseases in vitiligo cases from other populations have shown generally similar associations, although with perhaps some differences (Alissa et al., 2011, Ayanlowo et al., 2009, Chen et al., 2015, Gopal et al., 2007, Liu et al., 2005, Narita et al., 2011, Silva de Castro et al., 2012, Zhang et al., 2009) Likewise, these same autoimmune diseases occur at an elevated frequency in vitiligo probands’ first-degree relatives, suggesting that these autoimmune disease associations likely reflect at least partially shared genetic underpinnings (Alkhateeb et al., 2003).
This hypothesis has been completely borne out by genetic studies of vitiligo-associated genes and loci in other autoimmune diseases. Genetic associations of various major histocompatibility complex (MHC) genes and HLA alleles are of course well-established with many different autoimmune diseases. Of the 47 non-MHC loci that have been associated with vitiligo in the EUR population, six appear to involve genes that are specifically relevant to melanocytes and, thus, would not be expected to play roles in autoimmune diseases other than vitiligo. Of the remainder, at least 19 have also been genetically associated with at least one of the other autoimmune diseases that are epidemiologically associated with vitiligo (Figure 4). Many of these shared genetic associations appear to involve the same associated marker alleles and, thus, may potentially represent the same underlying causal variant. Thus, the hypothesis that shared epidemiological associations among various autoimmune diseases to some extent reflect shared underlying genetic predisposition has been fully confirmed.
Figure 4. Shared genetic associations among vitiligo-associated autoimmune diseases. In addition to various HLA alleles, at least 19 of the 47 non-MHC loci that have been associated with vitiligo in the EUR population have also been associated with at least one of the autoimmune diseases that are epidemiologically associated with vitiligo (autoimmune thyroid disease, adult-onset type 1 diabetes mellitus, rheumatoid arthritis, pernicious anemia, systemic lupus erythematosus, and Addison’s disease), in many cases involving the same associated marker alleles. Red, immune-related; blue, apoptosis-related; white, function unknown. Note that not all diseases have been studied to similar extents, and so for some diseases, relatively few genetic associations are yet known. EUR, European-derived white population; MHC, major histocompatibility complex.
Polygenic Architecture of Vitiligo Risk
As noted above, about 8% of vitiligo cases report at least one affected relative, with a pattern suggestive of non-Mendelian, complex inheritance (Laberge et al., 2005). Alkhateeb et al. (2003) found that the frequency of vitiligo in probands’ first-degree relatives was about 5–7%, depending on the ethnic population studied.
Roberts et al. (2019) combined data for the most significant SNPs from the 48 autosomal vitiligo GWAS loci—all relatively common—to create a vitiligo polygenic risk score, which they then used as a tool to probe various aspects of vitiligo genetic architecture in the EUR population. Comparing the top 1% of the cases with the rest, this risk score yielded an OR of 8.79—far higher than the risk scores for most other complex diseases (Khera et al., 2018)—owing at least in part to higher ORs of the vitiligo-associated variants than those of most complex diseases. Roberts et al. (2019) found that the positive predictive value of that vitiligo risk score was 71%, again better than polygenic risk scores for most other complex diseases (Wald and Old, 2019). Together, these results indicate that vitiligo is polygenic, although less so than many other complex diseases, and, as noted above, common genetic variants account for a high fraction (about 70%) of vitiligo heritability (Roberts et al., 2020a).
Roberts et al. (2019) also showed a major role for polygenicity in families with multiple affected relatives. They found that the vitiligo risk score in cases from such multiplex families was higher than that in singleton cases, with the risk score generally proportional to the number of affected relatives, and that this high polygenic risk was disproportionately transmitted to those affected relatives. Even in a family with a known high-risk rare variant (in FOXD3), the polygenic risk from common vitiligo variants was exceedingly high. Thus, whereas rare, high-penetrance variants undoubtedly play a role in some multiplex vitiligo families, even in such families, much of the genetic risk comes from over-representation of the same common risk variants involved in singleton vitiligo cases. This suggests that the application of a polygenic risk score might prove clinically useful for risk prediction in such families. Furthermore, the same genetic risk variants and corresponding biological pathways are involved in both simplex and multiplex vitiligo cases, suggesting that the same treatments will generally be applicable to both types of cases.
Vitiligo Triggering and Disease Onset
For autoimmune diseases such as vitiligo, case versus control genetic studies identify inherited risk factors that together increase the likelihood of loss of tolerance in response to an environmental trigger. Whereas no vitiligo environmental triggers are known with certainty, the Köbner phenomenon (Köbner, 1877) is particularly frequent in vitiligo (van Geel et al., 2011), suggesting a major role for skin damage in disease triggering.
Spritz et al. (2004) have studied the vitiligo age of onset in the EUR population as an indirect proxy for disease triggering. Jin et al. (2011) initially showed that vitiligo age of onset is genetically associated with the MHC class II region. These investigators (Jin et al., 2019) later showed that the distribution of vitiligo age of onset is bimodal and that early-onset vitiligo is genetically associated with a specific extreme-risk MHC class II haplotype (OR = 8.1) containing an enhancer variant that upregulates the expression of HLA-DQB1 mRNA and HLA-DQ protein on peripheral blood monocytes and dendritic cells. Elevated HLA class II protein expression on such professional antigen-presenting cells might thus enhance the response to triggering antigens, facilitating loss of tolerance by autoreactive T cells.
In addition to a genetic component underlying vitiligo triggering and onset, Jin et al. (2020) also searched for evidence for an environmental component by analyzing the long-term trends in vitiligo age of onset. Studying EUR vitiligo cases from North America and Europe, these investigators observed a dramatic shift over the period 1970–2004 from the mean onset of about 15 years of age in 1970 to over 30 years of age in 2004 (Figure 5). This change was unrelated to the extreme-risk MHC class II haplotype. Moreover, the pattern of change appeared generally similar among EUR vitiligo cases from both North America and Europe. Together, these findings suggest that exposure to one or more important vitiligo environmental triggers became reduced or delayed over this period in these populations. Whereas the cause of this seemingly beneficial change is not known, important clues might be gleaned by comparison with analogous data from other world populations.
Figure 5. Mean age at vitiligo onset in 4,406 EUR cases from North America and Europe by calendar year of onset, 1951–2013. Circles denote means, and vertical bars delimit 95% CIs. Black segments show regression lines for time periods of 1951–1969, 1970–2004, and 2005–2013. Reprinted from Jin et al. (2020) with permission. CI confidence interval; EUR, European-derived white population.
Perspectives and Future Directions
The past two decades have seen extraordinary progress in deciphering the genetic basis of vitiligo risk, both in general terms and in the identification of specific genes and genetic variants that underlie risk. This progress has provided a profound understanding of the pathobiology of vitiligo and its relationship to other autoimmune diseases. Vitiligo behaves as a complex polygenic disease. For the average vitiligo case, about 20% of risk comes from the environment, about 56% from various common genetic variants, and about 24% from a large diversity of rare variants. The common genetic variants associated with vitiligo risk, principally involved in immunoregulation, apoptosis, and melanocyte biology, can be combined in a vitiligo polygenic genetic risk score that has impressive positive predictive value compared with those for other complex diseases.
Nevertheless, many important questions remain. It is clear that the fundamental genetic architecture of vitiligo risk is generally polygenic and that vitiligo polygenic risk is generally additive at the macro level. However, it is not yet known whether polygenic vitiligo involves just one basic pathobiological process or there are multiple vitiligo biological endotypes, defined by different genes and pathways, with risk additive within each endotype. This is an important distinction, because the former would suggest that effective vitiligo treatments would be biologically generic, whereas the latter would suggest that different treatments might be needed for different vitiligo biological endotypes.
A related question is whether the general genetic architecture and pathobiology of vitiligo are similar in different world populations. Many vitiligo susceptibility loci and perhaps some corresponding causal genetic variants appear to be shared in patients from the EUR, CHN, and perhaps other populations. That would suggest that vitiligo in these populations involves shared pathobiological pathways. Nevertheless, other vitiligo susceptibility loci appear to be population-specific. That would suggest the possibility that some cases of vitiligo in one or another population might involve some aspects of differing underlying biology.
At least in the EUR population, about 70% of vitiligo genetic risk derives from common genetic variants. The remaining 30% of vitiligo genetic risk derives from a large diversity of rare genetic variants. As of now, only a single rare, highly penetrant vitiligo susceptibility variant, in FOXD3, has been identified, although it is likely that rare and uncommon variants in MC1R and IFIH1 also play roles. To identify additional specific rare vitiligo risk variants will likely require either association studies with much larger sample sizes or family-based studies to track the cosegregation of rare genomic variants along with genetic risk.
As of now, no large-scale genetic studies have yet addressed segmental vitiligo. The prevalence of associated autoimmune disorders does not appear to be elevated in segmental vitiligo (Iacovelli et al., 2005), suggesting that segmental vitiligo might not represent a fundamentally autoimmune process. Nevertheless, at least some susceptibility genes might be shared between segmental and nonsegmental vitiligo. Taïeb et al. (2008) speculated that segmental vitiligo might reflect some form of cutaneous somatic mosaicism. Might evidence of that be observed by single-cell DNA sequencing or other modern omics approaches?
Perhaps most important, all genetic studies of vitiligo to date have compared the genetic basis of case status with that of noncase status, either in large patient populations or within families. Fundamentally, such studies address the underlying biological determinants of disease occurrence; in the case of vitiligo, autoimmune triggering. Whereas that topic is of key importance, these biological determinants may not be the same as those that influence vitiligo clinical course once autoimmune triggering has occurred, which of course might have the greatest relevance to vitiligo treatment.
Finally, whereas there has thus been great progress toward discovering the genetic components of vitiligo susceptibility, as of now, no common environmental triggers for vitiligo have been identified with certainty. Such knowledge would provide an even deeper understanding of vitiligo pathobiology and furthermore, might provide opportunities to ameliorate vitiligo risk. Unfortunately, whereas we have a rigorous global systematic scientific method to identify genetic risk factors for disease, we have yet to develop an analogous global systematic approach to identify specific environmental risk factors.
Conflict of Interest
The authors state no conflict of interest.
Acknowledgement
We thank the thousands of patients with vitiligo, their family members, and normal control individuals around the world who participated in our studies of vitiligo genetics, as well as our many medical and scientific collaborators.
Richard A. Spritz1,2 and Stephanie A. Santorico
