Pathogenic variation in human DNA damage repair genes was originated from the evolutionary process of modern humans

Collecting genome sequences from ancient humans

It is well recognized that human genetic variation was largely arisen in recent human history.⁴⁶ However, it mainly refers to the biological adaptation-related, non-pathogenic variation. For the origin of DDR PVs in modern humans, it remains largely unclear. A comparison of genomic sequences between ancient humans and modern humans can provide clues to address the issue. However, it was technically impossible to do so before due to the lack of ancient human genomic sequences. As such, existing studies were mostly restricted to limited founder PVs in a few DDR genes through haplotyping⁴⁷ and molecular clock calculation.⁴⁸ However, the rapid development of ancient DNA extraction technologies in recent years has revolutionized the field. Genome sequences are available now from over 5000 ancient humans dated back to the latest human Out-of-Africa migration 50,000 to 60,000 years before present (BP) and the extinct hominins of Neanderthals and Denisovans.^49,50 For the first time, it is possible to trace the origin of modern human genetic variations in ancient humans.^51,52 For example, genetic variants in SLC22A4 predisposing high coronary heart disease risk were identified in a 5300 years old glacier mummy “Iceman Ötzi”,⁵³ genetic variants in EPAS1 enhancing human adaptation to hypoxia were identified in Denisovans,^53,54 and genetic traits for human susceptibility to COVID-19 infection and Dupuytren's Disease were identified in Neanderthals.^55,56 However, the studies have predominantly focused on genetic variations associated with adaptive changes in response to selective pressures from climate, lifestyle, and infectious diseases etc. No attempts have been made to address the origin of DDR PVs that predispose modern humans to cancer, the most devastating threats to human health.

We collected genome sequences from 5031 published ancient human genomes dated from 45,045 to 100 years BP, 12 Neanderthals dated from 120,000 to 38,515 years BP, and 4 Denisovans dated from 158,500 to 69,650 years BP.²¹ We then searched DDR PVs in the ancient humans. The quality of DNA extracted from ancient humans is critical to ensure the reliability of the data generated. Nowadays, the technologies for extracting DNA from ancient humans are highly advanced because of the outstanding work by the leading laboratories and scientists such as Svante Pääbo,⁵⁷ and standard protocols and strict extraction conditions are widely adapted to control the quality of the DNA extracted from ancient humans. The sources of DNA damage in ancient samples have been well determined, such as the exogenous microbe contamination in ancient samples and modern human DNA contamination during the process of DNA extraction; fragmentation of ancient DNA with breaks enriched next to purines; deaminated cytosine (C) to uracil (U) in ancient DNA caused mispairing during PCR amplification, resulting in C to T substitution in the amplified DNA; and oxidized pyrimidines and cross-links blocking amplification of the affected region in ancient DNA. Effective measures have been developed to maximally prevent the influence of these artifacts to ensure the integrity of the extracted DNA and the data derived from ancient DNA. For example, stringent laboratory protocols now cover the entire process of ancient DNA extraction, library preparation, DNA sequencing, and sequence analysis, such as the controlled laboratory environments for DNA extraction, amplification and indexing, and the use of uracil-DNA glycosylase with endonuclease VIII to limit uracil-driven mis-incorporation. Tight quality control during sequence analysis also plays crucial roles in further eliminating the artifacts, such as excluding terminal bases, eliminating duplicates and low-quality sequences, and removing the sequences from contaminated human mitochondria DNA and nuclear DNA. In our studies, we further applied two extra steps to ensure the reliability of DDR PVs derived from the ancient DNA sequences, including the use of binary-based, curated, quality-controlled BAM files rather than the raw text-based FASTQ read files, and the use of well-annotated DDR PVs in the ClinVar database as the references to determine their presence in ancient DNA rather than calling DDR PVs directly from ancient DNA sequences. These extra steps compensated for the inability of performing technical replicate validation due to the limitation of lacking ancient DNA samples.

Identifying DDR PVs conserved between modern and ancient humans

We searched the ancient humans for the presence of 1781 DDR PVs in 87 DDR genes identified in modern humans and validated them in ClinVar.⁴¹ We identified 1266 PVs in 73 DDR genes in 1019 ancient humans (Table 2; Table S1), accounting for 71% (1266/1781) of DDR PVs present in the modern human population. We constructed a database to host DDR PVs from ancient humans (https://genemutation.boboz.io/dbddr-AncientHumans).

The shared DDR PVs showed the following features:

• 1.

  Narrow range of arising time. The majority of ancient PV carriers were dated within the last 10,000 years BP (Fig. 3A) with an enriched range between 5000 and 1000 years BP, except for a few ancient carriers out of the range. The oldest PV was BRCA1 c.181T > G p.(Cys61Gly) present in a Russia carrier dated 37,470 years BP, and the youngest PVs were MSH2 c.1204C > T p.(Gln402∗), MUTYH c.437G > A p.(Trp146∗) and RNASEH2A c.69G > A p.(=) present in a Vanuatu carrier dated 190 years BP (Fig. 3B and Table S1);

• 2.

  Modern and ancient humans had similar loads of DDR PVs. Comparing with the 1,781 PVs in 81 DDR genes identified in global ethnic populations,⁵¹ we observed similar quantitative loads of DDR PVs between ancient humans and modern humans, except for the PVs in TP53, MLH1, and MSH2, which showed lower loads in modern humans than in ancient humans (Fig. 3C).

Identifying modern human DDR PVs sharing with Neanderthals and Denisovans

The human genome inherited 2%–4% of DNA from extinct Neanderthals and Denisovans by admixture.^58,59 This suggests a possibility that certain disease susceptibility in modern humans could be originated from the extinct hominins.60, 61, 62 Indeed, we identified a set of modern human DDR PVs in Neanderthals and Denisovans, including 4 TP53 PVs in three Neanderthals dated 44,000, 42,540 and 38,515 years BP, two TP53 PVs in a Denisovan dated 158,550 years BP, 3 ATM PVs, 1 BRCA2 PV and 1 CHEK2 PV in 5 Neanderthals dated 80,000 to 50,000 years BP (Fig. 4).²¹ These findings suggest that certain cancer predisposition alleles in modern humans were likely originated from Neanderthal and Denisovan.

Out of Africa migration was a major landmark for DDR PVs arising in modern humans

Our research supports the emergence of DDR PVs within human populations (Fig. 3A and B). However, the precise time for PV arising remains uncertain considering the long evolutionary history of humans. It is well determined that the genetic diversity of modern humans was strongly influenced by the latest Out-of-Africa migration.⁶³ We hypothesized that the Out-of-Africa migration could be a cutting point for the arising of DDR PVs in modern humans due to bottleneck effects, adaptation, and selection after the migration.

To test the hypothesis, we first identified 260 DDR PVs from 28,872 modern Africans and 974 DDR PVs from 222,342 modern non-Africans from public resources. We then used the 1,266 DDR PVs identified from the ancient humans as the intermediator to compare the two datasets. The results show differential sharing of ancient DDR PVs: 70% of ancient DDR PVs were shared with non-Africans only, 26.7% with both Africans and non-Africans, and 3.8% shared with Africans only, suggesting that the majority of DDR PVs in modern humans were closer to the ancient humans than modern Africans did (Fig. 5A). However, the results might be biased due to the different sample sizes between modern Africans (n = 28,872) and non-Africans (n = 222,342) used in the comparison. To test the possibility, we performed a rarefaction analysis by defining 15 targeted subgroups from n = 1,000 to n = 28,872 for the 222,342 non-Africans. For each subgroup, we randomly drew n individuals from 222,342, and repeated the process 1000 times. We then identified DDR PVs in each subgroup and searched for their presence in ancient humans. The results showed that the ancient-shared PVs in the non-Africans remained remarkedly higher (median = 49.5%, 95th percentile = 51.5%) than those in the Africans (17.3%, 95th percentile = 18.5%) (Fig. 5B), indicating that regardless of balanced or unbalanced sample sizes, DDR PVs in modern humans were closer to the ancient humans than modern Africans did. Therefore, the results support the conclusion that most of DDR PVs in modern humans arose after the Out-of-Africa human migration. Nevertheless, we acknowledge that the limited data from Africans may still influence the accuracy of the conclusion. Future studies incorporating more data from Africans, either ancient or modern, should refine the conclusion more precisely.

Human admixture promoted the transmission of DDR PVs between ethnic populations

The global exploration initiated in the15^th century by the Europeans caused one of the most extensive admixtures in modern humans.⁶³ For example, the global exploration by Portuguese in the 15th century generated more than 20 million admixture population worldwide nowadays with Portuguese heritage, as exemplified by the over 5 million Brazilian population carrying Portuguese heritage.⁶⁴ We hypothesized that disease susceptibility in certain ethnic populations could be transmitted to other populations through admixture. We tested our hypothesis by searching for the Portuguese BRCA founder PVs in the Brazilian population. From the 255 BRCA PVs identified in 7,711 Brazilians, 29 BRCA PVs were found to be originated from the Portuguese, including 14 BRCA1 Portuguese founder PVs, and 15 BRCA2 Portuguese founder PVs.⁴³ These results confirm that DDR PV-related cancer susceptibility is indeed transmittable through human admixture.

Arising timing of DDR PVs validated by haplotyping-determined DDR founder PVs

Our research revealed that human DDR PVs were mainly originated from human themselves in the past few thousand years. Historically, the arising time for many well-known DDR founder PVs has been determined via the haplotyping method.⁶⁵ We used the determined arising time of these DDR founder PVs as the references to validate the arising times of DDR PVs determined by our study. We collected the dated time from 68 founder PVs in seven DDR genes, including BRCA1, BRCA2, MLH1, MSH2, MSH6, MUTYH and TP53. The arising time for these founder PVs were all within the recent 4,000 years BP except MUTYH c.1103G > A dated 8,675 years BP and MUTYH c.452A > G dated 7,500 years BP. For example, in the three BRCA founder PVs in Ashkenazi Jewish, BRCA1 185delAG (c.68_69del) was 1,500–750 years BP,⁶⁶ BRCA1 5382insC (c.5266dup) was 1,800 years BP,⁶⁷ and BRCA2 6174delT (c.5946del) was 580 years BP⁶⁸; the BRCA1 c.4035delA in Balts was 1,550 years BP⁶⁹; BRCA2 c.771_775del in Icelander was 500 years BP⁷⁰; BRCA2 c.7480C > T and c.8327 T > G in Finnish were 400–200 BP and 220–140 years BP⁷¹; TP53 c.1010G > A in Brazilian was 2,000 years BP.⁷² The arising timing determined by haplotyping analysis is well within the scope dated by our anthological analysis, further supporting the short history of DDR PVs in modern humans (Table 3).