Repertoire Sequencing of B Cells Elucidates the Role of UNG and Mismatch Repair Proteins in Somatic Hypermutation in Humans

Abstract

The generation of high-affinity antibodies depends on somatic hypermutation (SHM). SHM is initiated by the activation-induced cytidine deaminase (AID), which generates uracil (U) lesions in the B-cell receptor (BCR) encoding genes. Error-prone processing of U lesions creates a typical spectrum of point mutations during SHM. The aim of this study was to determine the molecular mechanism of SHM in humans; currently available knowledge is limited by the number of mutations analyzed per patient. We collected a unique cohort of 10 well-defined patients with bi-allelic mutations in genes involved in base excision repair (BER) (UNG) or mismatch repair (MMR) (MSH2, MSH6, or PMS2) and are the first to present next-generation sequencing (NGS) data of the BCR, allowing us to study SHM extensively in humans. Analysis using ARGalaxy revealed selective skewing of SHM mutation patterns specific for each genetic defect, which are in line with the five-pathway model of SHM that was recently proposed based on mice data. However, trans-species comparison revealed differences in the role of PMS2 and MSH2 in strand targeting between mice and man. In conclusion, our results indicate a role for UNG, MSH2, MSH6, and PMS2 in the generation of SHM in humans comparable to their function in mice. However, we observed differences in strand targeting between humans and mice, emphasizing the importance of studying molecular mechanisms in a human setting. The here developed method combining NGS and ARGalaxy analysis of BCR mutation data forms the basis for efficient SHM analyses of other immune deficiencies.

Keywords: B cells; B-cell receptor; DNA repair; base excision repair (BER); constitutional mismatch repair deficiency (CMMRD); immunoglobulin; mismatch repair (MMR); somatic hypermutation.

Figure 1

Mutations in MMR and BER proteins lead to reduced frequency of SHM. (A) Median SHM frequency in IGHG and IGHA rearrangements is reduced in UNG-, MSH2-, MSH6-, and PMS2-deficient patients compared to HCs. Frequency of IGHG (B) and IGHA subclasses (C). Patients with UNG-, MSH2-, MSH6-, and PMS2-deficiency have more IGG1, IGG3, and IGA1 rearrangements compared to HCs. Statistical significance was performed using a Mann–Whitney test and indicated using ^*P < 0.05 and ^**P < 0.01.

Figure 2

UNG and MMR deficiency result in changes in SHM patterns. (A) Transition tables of HCs, UNG-, MSH2-, MSH6-, and PMS2-deficient patients. The percentage of mutations at GC base pairs that are transversions (B) and the percentage of mutations at AT base pairs (C) in HCs and UNG-, MSH2-, MSH6-, and PMS2-deficient patients. The percentage of SHM present in WA/TW (D) and RGYW/WRCY motifs (E) in HCs and patients with genetic defects in MMR or BER. (F) Absolute frequency of mutations at AT base pairs and transitions and transversions at GC base pairs as compared to age-matched controls (n = 5 for age 3–4 and n = 10 for age 6–22). Statistical significance was performed using a Mann–Whitney test and indicated using ^*P < 0.05, ^**P < 0.01, and ^***P < 0.001.

Figure 3

Defects in UNG, MSH2, MSH6, and PMS2 influence the corrected A/T ratio. (A) Transition tables of the percentage of mutations corrected for the A,C,T, and G content of the sequenced transcripts. (B) The corrected A/T ratio in HCs and patients with genetic defects in UNG, MSH2, MSH6, and PMS2. Statistical significance was performed using a Mann–Whitney test and indicated using ^**P < 0.01.

Figure 4

Minor differences are found in SHM patterns between mice and human. Differences in the percentage of GC mutations that are transversions (A), the percentage mutations at AT base pairs (B), and the A/T ratio (C) between mice and human with and without genetic defects in genes involved in BER or MMR. Statistical significance was performed using a Mann–Whitney test and indicated using ^*P < 0.05.

Figure 5

Model of the molecular mechanism of SHM in human. SHM is initiated by AID, which creates a U:G mismatch and can lead to mutations via five different routes. (I) During replication, the U can be recognized as a T leading to C>T and G>A transversions. (II) The U can be recognized and removed by UNG, creating an AP site. During subsequent cell division, this AP site is non-instructive and therefore TLS polymerases incorporate a random nucleotide at this position leading to transition and transversions at GC base pairs. (III) A second UNG-dependent pathway involving POLH can be initiated and lead to mutations at AT base pairs. (IV) If the U:G mismatch is recognized by the MSH2/MSH6 dimer, multiple bases surrounding the U:G mismatch can be removed by EXO1, creating a single-strand gap. Next, PCNA is polyubiquitinated and recruits POLH to fill this gap, thereby creating mutations at AT base pairs. (V) Finally, the MSH2/MSH6-, UNG-hybrid pathway can be initiated with two U's that are present in close proximity to each other but on opposing strands. Here, one of the U's is recognized by MSH2/MSH6 and processed to create a single-strand gap. The other U is processed by UNG, thereby creating an AP site opposite of the single-stranded gap. Next, when the single-stranded gap is filled, a TLS polymerase is recruited to fill the gap opposing the AP site, thereby leading to MSH2/MSH6-dependent transversions at GC base pairs.