Background

Modern humans originated from Africa 100,000 - 60,000 years ago and migrated to different global destinations [1]. Genetic variation in different environments has resulted in differential variation signatures between different ethnic populations. During the process of human evolutionary history, admixture between geographically isolated populations occurred by means of migration, occupation, displacement, and war etc. [2]. However, the global colonization by Europeans in recent hundreds of years has resulted in the most widespread admixture in human history [3]. Biomedically, admixture could not only enhance genetic diversity, but may also transfer disease susceptibility between populations [4, 5]. The impact of admixture on human health gains increased attention [6, 7]. For example, modern humans inherited multiple loci in chromosome 3 from the extinct Neanderthals, which can regulate the expression of CCR1 and CCR5, the critical chemokine receptor genes involving the severe immune response to COVID-19 infection [8, 9]; the pathogens carried by Europeans to the New World caused catastrophic consequence on Native Americans [10]. However, most of the previous studies focused on the susceptibility for infectious diseases. Little is known for the impact of admixture on susceptibility for chronic diseases, such as cancer, in modern humans.

The Portuguese global exploration provides a typical example of human admixture. Initiated in the 15th century, the Portuguese established its colonies across Africa, Asia and America, including countries of Angola, Mozambique, India, Sri Lanka, Malaya, China, Japan, East Timor and Brazil etc. [11] (Fig. 1). In many of the places colonized by Portuguese, admixture formed between Portuguese and locals. Up to now, there are over 20-million Portuguese-heritage population worldwide in comparing to 10.3-million native Portuguese in Portugal [12] (Table S1). For example, around five-million Brazilian individuals contain Portugal heritage constituting 2.3% of Brazil population, over 8,000 Macau individuals named Macanese contain Portugal heritage constituting 1.5% of Macau population [13]. The population with Portuguese heritage provides a valuable model to study the impact of admixture on human health.

Fig. 1
figure 1

Portuguese global exploration since the 15th century. It outlines the East path towards Asia and the West path towards South America and the Pacific. Red: Portuguese colonized countries and regions

Full size image

Genomic DNA in the genome is constantly damaged by external and internal factors. The damaged DNA must be repaired in order to prevent genome instability, which otherwise can cause high susceptibility to diseases, in particular cancers. This function is achieved by the DNA damage repair (DDR) system consisting of multiple pathways and hundreds of different genes. The pathways include homologous recombination (HR) and nonhomologous end joining (NHEJ) pathways repairing double strand DNA breaks; base excision repair (BER) pathway repairing small, non-helix–distorting base lesions; direct reversal (DR) repair pathway repairing the DNA damaged by ubiquitous alkylating agents; Fanconi anemia (FA) pathway repairing strand cross-link errors; mismatch repair (MMR) pathway repairing mismatch errors; and nucleotide excision repair (NER) pathway repairing helix-distorting DNA lesions. The joint activities of a group of genes in each DDR pathway repair the given type of DNA damage [14]. For example, in homology recombination pathway, BRCA1 interacts with PALB2 and BRCA2/RAD51 in the double strand break sites to initiate the repair process, and BRCA2 interacts with RAD51 to single-stranded DNA to promote DNA strand exchange and homologous pairing to repair the damage DNA [15, 16].

However, DDR genes themselves are prone to genetic variation. The variation with pathogenic effects can damage the function of the affected genes, leading to disease susceptibility. For example, women carrying pathogenic variants (PVs) in BRCA1 have a 55%–72% lifetime risk of developing breast cancer, 39%–44% lifetime risk of developing ovarian cancer; women carrying pathogenic variants in BRCA2 have a 45%–69% risk of developing breast cancer and 11%–17% lifetime risk of developing ovarian cancer [17,18,19]. PVs in BRCA1 and BRCA2 also increase the risk of developing prostate cancer, colorectal cancer and pancreatic cancer [20]. Identification of BRCA PV carriers without cancer can prevent cancer development by taking preventive measures through early cancer surveillance, chemoprevention and preventive surgery [21,22,23,24,25], and identification of BRCA PV carriers with cancer can benefit the use of targeted therapy, such as PARP inhibitors, to treat the cancer [26]. In our current study, we targeted BRCA1 and BRCA2 (BRCA) based on the following considerations: 1). High clinical significance. It is well determined that BRCA PV is highly linked to cancer risk, and is widely used in clinical practice as the genetic marker for cancer diagnosis, prevention, prognosis, and treatment [27,28,29,30,31]. Therefore, information from BRCA study could be highly significance for clinical applications; 2). High prevalence. among all DDR genes, BRCA PV has the highest prevalence that BRCA1 PVs reaches 0.079% and BRCA2 PVs 0.124% in worldwide human populations, whereas the prevalence of the PVs other DDR genes was much lower, e.g., 0.034% in CHEK2, 0.015% in TP53, 0.010% in RAD51D, and 0.002% in POLK [32]. The higher prevalence of BRCA PVs implies higher relevance of the results to human health; 3). Availability of rich variation data. BRCA was among the first determined cancer risk genes [33]. Extensive efforts made in the past decades have identiifed over 70,000 genetic variants in human BRCA [34] (https://brcaexchange.org/factsheet). Importantly, BRCA variation has also been extensively characterized in both Portuguese population and Portuguese heritage-rich Brazilian population.

In current study, we used the BRCA PVs from Portuguese and Brazilian populations as a model to study the impact of admixture on cancer susceptibility. By comprehensive collection, standardization and classification of the BRCA variation data originated from Portuguese and Brazilian populations, we generated the centralized BRCA variation data in both Portuguese and Brazilian populations. By comparing the BRCA PV data between Portuguese and Brazilian populations, we observed the 0 presence of Portuguese-originated BRCA PVs variation in Brazilian population. Our study provides evidence showing the impact of admixture on human cancer susceptibility.

Results

BRCA variation in Portuguese population

We first generated a comprehensive collection of Portuguese-originated BRCA variation data. We identified a total of 23 publications between 2000 and 2022 reporting the BRCA variation data in 11,482 Portuguese cancer patients, mostly breast cancer, ovarian cancer, prostatic cancer, and HBOC (Hereditary Breast and Ovarian Cancer Syndrome), largely reported by the Laboratory of Dr. Manual Teixeira in Portuguese Oncology Institute of Porto, Portugal ([35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57], Table S1). We extracted the BRCA variation data from the publications. Upon standardization, annotation and integration, we identified a total of 220 non-redundant BRCA variants, including 93 (42.3%) in BRCA1 and 127 (57.7%) in BRCA2. By referring to ClinVar database, the 220 BRCA variants were classified into different clinical classes, including 78 (35.5%) Pathogenic or Likely Pathogenic variants (PV), 66 (30.0%) Benign and Likely Benign variants (BV), 55 (25%) Variant of Unknown Significance (VUS), 24 (10.9%) conflict of interpretation, and 30 (13.6%) unclassified variants (Table 1A, Fig. 2A, B, Table S2). Single base variation accounted for the majority of variants (65.4%) (Table S5). In the 11,482 cancer patients tested in the original studies, 556 (4.8%) cases were identified as the carriers for the 78 BRCA PVs (Fig. 2B). We also identified 8 large structural variation, 5 in BRCA1 of one deletion covering NRB1 exon 1 to exon 11 and BRCA1 exon 1 to exon 7, one duplication covering exon 1 to exon 3, one duplication covering exon 4 to exon 6, one deletion affecting exon 11 and one deletion affecting exon 13; and three in BRCA2 of one insertion in exon 3, one duplication from exon 17 and exon 18, and one duplication in exon 21 (Fig. 2C, Table S3, Table S4).

Table 1 BRCA variants identified in Portuguese
Full size table
Fig. 2
figure 2

Features of Portuguese BRCA variation. A Clinical classification of BRCA variants. P: Pathogenic; PLP: Pathogenic or Likely Pathogenic; LP: Likely Pathogenic; B: Benign; LB: Likely Benign; BLB: Benign or Likely Benign; VUS: Variants of Uncertain Significance; CIP: Conflicted Interpretation of Pathogenicity; U: Unknown. B BRCA PVs identified in Portuguese cancer cohort. The left pie showed 4.8% of 11,482 Portuguese cancer patients were the BRCA PV carriers; the right pie showed that 59.5% of the PV carriers carried the three Portuguese founder pathogenic variants, and 40.5% carried non-founder pathogenic variants. C Large structural variants in Portuguese BRCA variation

Full size image

Portuguese BRCA founder variants

Founder variant is defined as “a pathogenic variant observed at high frequency in a specific population due to the presence of the variant in a single ancestor or small number of ancestors [58]. Three BRCA PVs are well determined as Portuguese BRCA founder variants, including BRCA1 c.2037delinsCC, BRCA1 c.3331_3334del, and BRCA2 c.156_157insAlu [59]. Of the 556 BRCA PV carriers identified in the 11,482 cancer patients, 331(59.5%) were the carriers for the three founder variants, including 24 (4.3%) carriers of BRCA1 c.2037delinsCC, 58 (10.4%) carriers of BRCA1 c.3331_3334del, and 249 (44.8%) carriers of BRCA2 c.156_157insAlu (Fig. 2B).

Portuguese-specific BRCA PVs

We searched the Portuguese-originated BRCA PVs in GnomAD database v.3.1.2, which contained genetic variation data from 76,156 individual genomes of African, African American, Amish, Ashkenazi Jewish, East Asian, Middle Eastern, South Asian, Latino, Admixed American, Finnish European, and non-Finnish European. We observed that 56 (71.8%) of the 78 Portuguese BRCA PVs were not present in non-Portuguese populations (Table 1B), the remaining 23 PVs (29.1%), they were mainly shared with the non-Finnish European populations (Table S5).

BRCA variation in Brazilian population

We identified 24 publications reporting BRCA variation in 7,711 Brazilian individuals mostly breast cancer, ovarian cancer and HBOC patients (Table S6A). We extracted the BRCA variants from the reports, and performed standardization, annotation, and classification. We identified a total of 255 BRCA PVs, 136 (53.3%) in BRCA1 with 743 carriers and 119 (46.6%) in BRCA2 with 365 carriers (Table S6B, Table S6C). The carrier rate of BRCA PV was about 3.3% in the tested 7,711 Brazilian cancer patients.

BRCA variation databases

We developed two open-access databases, one is the dbBRCA-Portuguese to host the Portuguese-originated BRCA variation data (https://genemutation.fhs.um.edu.mo/dbbrca-portuguese/) and the other is the dbBRCA-Brazilian to host the Brazilian-originated BRCA variation data (https://genemutation.fhs.um.edu.mo/dbbrca-brazilian/). The databases provide the information on variants, annotation, clinical classification, cancer type, carrier frequency, and original references etc., users can perform online search or download the entire datasets for their own analysis.

Presence of Portuguese BRCA PVs in Brazilian population

Brazil has a population size of over 215 million, of which 5 million (2.3%) has Portuguese-heritage background (https://data.worldbank.org/indicator/SP.POP.TOTL?locations=BR). We tested if the Portuguese BRCA PVs could be present in Brazilian population. To ensure high reliability of the results, we used only the 56 Portuguese-specific BRCA PVs for the comparison. We observed that 29 (51.8%) of the 56 Portuguese-specific BRCA PVs were present in Brazilian population with 254 carriers, of which 119 (46.7%) carried the three Portuguese founder variants, including 14 in BRCA1 including the Portuguese founder BRCA1 c.3331_3334del [84 carriers in 4,131 (2.03%) tested individuals], BRCA1 c.2037delinsCC [5 carriers in 1,261 (0.4%) tested individuals], and 15 in BRCA2 including the Portuguese founder BRCA2 c.156_157insAlu [30 in 4,471 (0.67%) tested individuals]. There were 254 carriers in the 7,711 tested Brazilian cancer patients, reaching to the carrier frequency of 3.2% (Fig. 3A, Table S6B, Table S6C). Searching the shared PVs in Global Minor Allele Frequency (GMAF) and Genome Aggregation Databases (gnomAD) via ClinVar showed that these shared variants had either no or low population frequency information, indicating that they were all rare variants (Fig. 3A). None of the eight large BRCA structural variations in Portuguese were identified in Brazilian. The Brazilian carriers for the 29 shared BRCA PVs were residents in 13 of the 26 States and 1 Federal District of Brazil (Fig. 3B).

Fig. 3
figure 3

Portuguese-specific BRCA PVs in Brazilian population. A 29 of the 56 Portuguese-specific BRCA PVs identified in Brazilian population; B Resident location of the Brazilian carriers of Portuguese BRCA PVs in the states of Brazil (Darker blue)

Full size image

Our analysis of GenomAD database showed that the Portuguese-derived BRCA PVs were highly Portuguese-Specific, and widely present in Brazilian population. However, native American, African and European etc. all contribute to the genetic background for today’s Brazilian [60]. Besides the BRCA PVs in Brazilian originated from Portuguese (European), possibility exists that certain BRCA PVs might also be inherited from Native American or African. If this could be the case, assigning Portuguese as the sole source of BRCA PVs would not be appropriate. BRCA variation has not been systematically tested in Native American, which restricts the analysis for the relationship of BRCA PVs between native American and Brazilian. However, BRCA variation data from African are available. This allows us to test the relationship between Portuguese, Brazilian and African. We compared the BRCA PVs between African, Brazilian, and Portuguese. The PVs included the 134 African BRCA PVs (61 in BRCA1 and 73 in BRCA2) identified by CIMBA (the Consortium of Investigators of Modifiers of BRCA1/2) study [61], the 255 BRCA PVs in Brazilian, and the 78 BRCA PVs in Portuguese. The results showed that only 1 of the 136 (0.7%) Brazilian BRCA1 PVs (c.2389_2390del), 5 of the 119 (4.2%) Brazilian BRCA2 PVs (c.93G>A, 5197_5198del, c.5616_5620del, c.7900del, c.8009C>T) were present in African, and only 1 of the 119 (0.8%) Brazilian BRCA2 PVs (c.2808_2811del) was present in all three populations (Fig. 4). None of the 7 shared PVs were the Portuguese founder PVs nor within the 56 Portuguese-specific BRCA PVs defined after filtering in GenomAD database (Fig. 1B). The results ruled out the possibility that African contributed to the BRCA PVs shared between Portuguese and Brazilian.

Fig. 4
figure 4

Comparison of the BRCA PVs between African, Brazilian, and Portuguese. The PVs included 134 African BRCA PVs, 255 BRCA PVs in Brazilian, and 78 BRCA PVs in Portuguese. Only 1 of the 136 Brazilian BRCA1 PVs and 5 of the 119 Brazilian BRCA2 PVs were present in African, and 1 of the 119 Brazilian BRCA2 PVs was present in all three populations. The shared PVs were not Portuguese founder PVs, and not included in the 56 Portuguese-specific BRCA PVs

Full size image

We also searched the evolution origin for the 78 BRCA PVs in Portuguese. From the genomic data derived from over 5,000 ancient humans, we identified only 8 (10%) of the 78 BRCA PVs (3 in BRCA1 and 5 in BRCA2) present in ancient Europeans dated between 37,470 to 38,18 years before present, of which only BRCA2 c.9004G>A, c.8488-1G>A and c.7007G>A were among the 56 Portuguese-specific BRCA PVs (Table 2). Although the 8 Portuguese PVs highlighted their European origin, the fact that 70 of the 78 BRCA PVs were not present in the ancient humans further confirmed their Portuguese-specificity.

Table 2 Evolutionary origin of the BRCA PVs in Portuguese
Full size table

Discussion

Portuguese played a unique role in modern human admixture through its multi-century’s global colonization across Africa, Asia and America. The global distribution of Portuguese genetics provides an attract model to study the impact of admixture on human health. BRCA PVs are well determined as the genetic predisposition for high cancer risk. They provide valuable genetic markers to trace the transmission of pathogenic variation between human populations by admixture. By tracing the distribution of BRCA PVs originated in Portuguese in the Portuguese heritage-enriched Brazilian population, we observed high sharing of Portuguese BRCA PVs in Brazilian population, as represented by the Portuguese founder PV of BRCA1 c.2037delinsCC, BRCA1 c.3331_3334del, and BRCA2 c.156_157insAlu.

There are three models in explaining the genetic predisposition contributing to disease risk. Model 1: common variants contributing to common disease risk; model 2: low-frequent variants contributing to intermediate disease risk; and model 3: rare-frequent variants contributing to high disease risk [62]. For the BRCA PVs shared between Portuguese and Brazilian populations, they were either no or at extremely low frequencies of 0.00001 as reported in GnomAD (Fig. 3A). Therefore, these shared BRCA PVs are all rare variants, and the Model 3 fits the best to explain the cancer risk caused by these shared PVs. Genetic factor contributes 5–10% of breast cancer risk. Around 50% of the genetic factor is currently known that about 30% is by the rare BRCA PVs [63], 10-15% by the low frequent PVs in other genes such as PTEN, TP53, CHEK2, PALB2 and STK11 [64], and the 1-5% by the common variants. The rare and low frequent variants are often identified in cancer patients, the common variants are often identified in other genes by GWAS analysis at the population level. The rare variants in BRCA and low frequency variants in other genes often have higher penetrant in causing high cancer risk and high values for clinical applications, whereas the common variants often have low penetrant in causing low disease risk, which are often used for theoretical explanation of disease risk but not used for clinical applications. Nearly all BRCA PV-caused breast cancer is monogenic, implying that BRCA PV alone is sufficient to cause high cancer risk. The rarity of the BRCA PVs has an important clinical implication that individual-based approaches need to be taken to identify the PV carriers.

The three Portuguese-BRCA founder PVs of BRCA1 c.2037delinsCC, BRCA1 c.3331_3334del, and BRCA2 c.156_157insAlu contribute nearly 60% of Portuguese BRCA PV carriers (Table 1B). The situation is similar to the three founder PVs of BRCA1 c.68_69delAG, c.5266dupC, and BRCA2 c.5946delT in 1 of 40 Ashkenazi Jews [65], and the three BRCA1 founder PVs of C61G, 4153delA and 5382insC for 80% of the pathogenic BRCA1 PV carriers in Polish population [66]. Although BRCA1 c.3331_3334del is a founder variants in Portuguese, it is also present in the Iberia population [67]. Therefore, attention needs to pay in interpreting BRCA1 c.3331_3334del as Portuguese-specific. BRCA2 c.156_157insAlu accounted 46% of BRCA PVs in the Portuguese cancer cohort in our current study, much higher than the originally reported one-fourth, 33.6% or 37.9% [68,69,70]. This is likely due to the larger size of the combined patients than those in the individual original studies. Further, BRCA2 c.156_157insAlu cannot be detected by regular PCR-sequencing approach but requires the use of two independent PCR with specific primers for sequencing. This feature suggests that the actual prevalence of BRCA2 c.156_157insAlu could be higher than currently known. The presence of the three Portuguese founder variants in Brazilian provides strong evidence for their Portuguese origin. For other shared variants, some could occur coincidently in Portuguese and Brazilian, although the possibility may not be high considering their highly Portuguese-specific nature.

Ethnic specificity of PVs in cancer predisposition genes reflects genetic diversity and environmental adaptation of different ethnic populations, and can contribute to different susceptibility to disease risk in different populations [32, 71,72,73]. Ethnic specificity is contributed by the factors of evolution selection, bottleneck effects, genetic drift, and founder variation etc. during human evolution. Specifically, human BRCA PVs mostly arose in recent thousand years before present [74], and modern admixture initiated mostly in only a few hundreds of years ago. Referring to the BRCA PVs in a donor ethnic population should facilitate the identification of the PV carriers in the recipient population.

In conclusion, our study provides evidence that admixture can indeed impact on human health, in this case the cancer susceptibility, through transferring pathogenic variation from donor population to recipient population. While our study tested only BRCA PVs, It would be interesting to know if similar situation could be present in other disease risk genes.

Materials and methods

Sources and annotation of BRCA variation data in Portuguese and Brazilian populations

We performed a comprehensive literature search in PubMed and Google Scholar by using the terms “BRCA1”, “BRCA2”, “Portugal”, “Portuguese”, “Brazil”, “Brazilian” “variant”, and “mutation” in English and Portuguese, to identify the sources reporting BRCA variants originated from Portuguese and Brazilian population. We extracted the BRCA variation information from the original sources, including cDNA variant designation, amino acid designation, variant type, tested sample size, carrier number and cancer types. The collected BRCA variants were standardized following the human genome variation society (HGVS) nomenclature using the following references sequences: human genome reference: GRCh38.14(GCF_000001405.40); BRCA1: HGVSg: NC_000017.11, HGVSc: NM_007294.4, HGVSp: NP_009225.1; BRCA2: HGVSg: NC_000013.11, HGVSc: NM_000059.4, HGVSp: NP_000050.2. Variants were annotated by using ANNOVAR. Clinical class of the variants was based on ClinVar classification [75]. The carrier numbers for the same variants of different studies were combined to calculate carrier frequency in the tested cases. "1" was assigned for the variants without carrier information in the original sources. GnomAD v.3.1.2 database was used to search for variation frequency and Portuguese-specific BRCA PVs [76].

Database construction

Two open-access databases, dbBRCA-Portuguese and dbBRCA-Brazilian, were constructed to host the BRCA variant data from Portuguese and Brazilian populations following the procedures [18]. Briefly, the databases were designed using the Linux operating system (CentOS 7, (https://www.centos.org/)). User searching and retrieval requests were performed by MySQL relational database management (5.6.50, (https://www.mysql.com/)) and PHP (version 7.3, (https://www.php.net/)). Web front-end languages (HTML, CSS, and JavaScript) were used to implement a user interface for the databases.

Ancient human genomic data analysis

The process followed the procedures [77]. Briefly, ancient human genome sequences were downloaded from the Allen Ancient DNA Resource including 5,047 ancient humans dated between 45,045 and 100 years before present (BP) (version 50.0, https://reich.hms.harvard.edu/allen-ancient-dna-resource-aadr-downloadable-genotypes-present-day-and-ancient-dna-data, accessed October 19, 2021). BRCA variants in ancient humans were called by SAMtool and annotated by ANNOVAR, and compared with the 78 BRCA PVs shared between Portuguese and Brazilian. The location and dated age of ancient human individuals carrying the matched BRCA PVs were extracted from the original data.