To the Editor: Frameshift neoantigen-based cancer prevention vaccines (e.g., Nous-209, NCT05078866) are being tested in clinical trials for patients with Lynch syndrome (LS) caused by monoallelic mismatch repair (MMR) gene pathogenic variants (PVs). A liquid biopsy biomarker panel has been developed for profiling frameshift mutations (FSMs) in tumors and plasma cell-free DNA (cfDNA) for disease surveillance of LS (1). However, it is unknown whether the same vaccine strategy and biomarker panel can be applied to patients with constitutional mismatch repair deficiency (CMMRD) with germline biallelic MMR gene PVs. CMMRD is associated with an increased risk of developing brain tumors with high frequency (51%) and has the poorest outcomes among other malignancies, starting in childhood (2). We hypothesized that those loci mutated in LS would also be frequently mutated in CMMRD malignancies and that the vaccines and biomarker panel developed for LS would be applicable to CMMRD.

To test this hypothesis, 2 patients with CMMRD with brain tumors and biallelic germline PMS2 mutations were studied (Supplemental Table 1; supplemental material available online with this article; https://doi.org/10.1172/JCI185999DS1). Specimens were collected at the time of residual tumor resection for patient 1, at 6 years of age. For patient 2, who had glioblastoma (GBM), tumor tissue was collected at age 23, a year after blood collection.

To determine whether FSMs discovered in LS were also present in CMMRD, tumor DNA and cfDNA from 2 patients with CMMRD were sequenced using a 122-gene panel designed around loci known to be frequently mutated in LS tumors and loci corresponding to frameshift neoantigens in vaccines. A high number of FSMs were detected in tumors (Table 1 and Supplemental Table 2) (n = 214 [80% of genes in the panel; patient 1] vs. n = 52 [32% of genes in the panel; patient 2]), although patient 2 was older at the tumor collection (23 years of age) and had developed other cancers prior to GBM diagnosis. Deep allele frequency (DAF) was variable, with the highest DAF being 0.3592 for the ASXL1 A/AG variant (Supplemental Table 3). The discrepancy in the number of FSMs detected might be due to invasive residual tumor edge tissue used for patient 1 and/or clonal evolution of tumor cells in the mass (thus, the low DAF for most FSMs [below the detection level]) for patient 2.

Number of FSMs detected in matched brain tumor and cfDNA using a 122-gene panel

It is well accepted that brain tumors usually release fewer tumor fragments into the circulation owing to the blood-brain barrier. However, we observed a high number of FSMs in the blood (n = 43 [patient 1] vs. n = 47 [patient 2]) (Table 1). It is worth noting that the blood used for patient 2 was collected 1 year prior to GBM diagnosis, suggesting that FSMs may accumulate as a result of previously diagnosed tumors and/or FSMs could be detected before cancer diagnosis. Consistent with the finding from LS cfDNA (1), the number of FSMs detected in cfDNA was lower than that in tumors, especially for patient 1. The profile of FSMs was similar to what has been previously reported (3, 4) and what we have observed in LS (1). Moreover, only 1 FSM in cfDNA had a DAF of >0.05 (ASTE2 C/CT; DAF = 0.076; Supplemental Table 3), suggesting that FSMs were not enriched in cfDNA. DAF at individual loci was characteristic, and sequence context may be an important determinant.

Further analysis revealed that 94% and 42% of genes detected in cfDNA of patients 1 and 2, respectively, were shared with the paired tumor samples (Table 1). The low DAF in cfDNA may partially explain the discordance between the number of FSMs detected in blood and that detected in tumors. In addition, tumor heterogeneity, previous tumors in patient 2, and the likelihood of DNA fragments shed by all body cells in patients with CMMRD starting at a very young age may also account for the discordance.

To determine whether the number of FSMs detected in patients with CMMRD was higher than that in patients with LS, tumors from 3 young patients with LS with 2 paired cfDNA (see Patients and specimens in Supplemental Methods) were also sequenced using the same platform (Supplemental Table 4). The number of FSMs was much lower in LS tumors and cfDNA (n = 9–33 [7%–21% of genes in the panel] vs. n = 8–10 [7%–8% of genes in the panel], respectively), despite the presence of MMR gene loss of heterozygosity in the LS tumors. This indicates that FSMs may emerge and accumulate at a very young age from all body cells of patients with CMMRD and more DNA fragments may be shed and accumulated in the blood of patients with CMMRD than patients with LS.

It is challenging to care for patients with CMMRD because multiple primary cancers may emerge over the life course and affect a variety of different tissues. This increases the complexity of disease surveillance and management (5), which highlights the need for enhanced preventive and surveillance strategies. Our study provides evidence of the existence of FSMs in CMMRD and suggests that FSMs are not syndrome or cancer type specific. Further preclinical and clinical evaluation will be required to establish whether frameshift neoantigen-based vaccines and the biomarker panel developed for LS are effective in CMMRD.

Supplemental material

View Supplemental data

Footnotes

Conflict of interest: The authors have declared that no conflict of interest exists.

Copyright: © 2024, Song et al. This is an open access article published under the terms of the Creative Commons Attribution 4.0 International License.

Submitted: August 14, 2024; Accepted: October 15, 2024; Published: October 22, 2024.

Reference information: J Clin Invest. 2024;134(24):e185999.

References

1. Song Y, et al. Frameshift mutations in peripheral blood as a biomarker for surveillance of Lynch syndrome. J Natl Cancer Inst. 2024;116(6):957–965.View this article via: CrossRefPubMedGoogle Scholar
2. Ercan AB, et al. Clinical and biological landscape of constitutional mismatch-repair deficiency syndrome: an international replication repair deficiency consortium cohort study. Lancet Oncol. 2024;25(5):668–682.View this article via: CrossRefPubMedGoogle Scholar
3. Das A, et al. Genomic predictors of response to PD-1 inhibition in children with germline DNA replication repair deficiency. Nat Med. 2022;28(1):125–135.View this article via: CrossRefPubMedGoogle Scholar
4. Bajwa-Ten Broeke SW, et al. The coding microsatellite mutation profile of PMS2-deficient colorectal cancer. Exp Mol Pathol. 2021;122:104668. View this article via: CrossRefPubMedGoogle Scholar
5. Chung J, et al. DNA Polymerase and mismatch repair exert distinct microsatellite instability signatures in normal and malignant human cells. Cancer Discov. 2021;11(5):1176–1191.View this article via: CrossRefPubMedGoogle Scholar

致编辑：基于移码新抗原的癌症预防疫苗（例如 Nous-209、NCT05078866）正在对由单等位基因错配修复 （MMR） 基因致病性变异 （PV） 引起的林奇综合征 （LS） 患者进行临床试验。已经开发了一种液体活检生物标志物面板，用于分析肿瘤中的移码突变 （FSM） 和浆细胞游离 DNA （cfDNA），用于 LS 的疾病监测 （1）。然而，尚不清楚相同的疫苗策略和生物标志物组是否可以应用于具有种系双等位基因 MMR 基因 PV 的体质错配修复缺陷 （CMMRD） 患者。CMMRD 与高频脑肿瘤的风险增加 （51%） 相关，并且在其他恶性肿瘤中预后最差，从儿童期开始 （2).我们假设在 LS 中突变的那些基因座在 CMMRD 恶性肿瘤中也会经常发生突变，并且为 LS 开发的疫苗和生物标志物面板将适用于 CMMRD。

为了检验这一假设，研究了 2 例患有脑肿瘤和双等位基因种系 PMS2 突变的 CMMRD 患者（补充表 1;本文在线提供的补充材料;https://doi.org/10.1172/JCI185999DS1）。在患者 1 切除残留肿瘤时收集标本，年龄为 6 岁。对于患有胶质母细胞瘤 （GBM） 的患者 2，在采血一年后，在 23 岁时收集肿瘤组织。

为了确定在 LS 中发现的 FSMs 是否也存在于 CMMRD 中，使用 122 基因面板对来自 2 名 CMMRD 患者的肿瘤 DNA 和 cfDNA 进行测序，该面板围绕已知在 LS 肿瘤中经常突变的基因座和对应于疫苗中移码新抗原的基因座设计。在肿瘤中检测到大量 FSM（表 1 和补充表 2）（n = 214 [面板中 80% 的基因;患者 1] vs. n = 52 [面板中 32% 的基因;患者 2]），尽管患者 2 在肿瘤收集时年龄较大（23 岁），并且在 GBM 诊断之前已经患上其他癌症。深度等位基因频率 （DAF） 是可变的，ASXL1 A/AG 变体的最高 DAF 为 0.3592（补充表 3）。检测到的 FSM 数量的差异可能是由于患者 1 使用的侵袭性残留肿瘤边缘组织和/或患者 2 的肿块中肿瘤细胞的克隆进化（因此，大多数 FSM 的低 DAF [低于检测水平]）。

使用 122 基因面板在匹配的脑肿瘤和 cfDNA 中检测到的 FSM 数量

众所周知，由于血脑屏障，脑肿瘤通常会向循环中释放较少的肿瘤碎片。然而，我们观察到血液中存在大量 FSM （n = 43 [患者 1] vs. n = 47 [患者 2]）（表 1）。值得注意的是，患者 2 使用的血液是在 GBM 诊断前 1 年采集的，这表明 FSM 可能由于先前诊断的肿瘤而积累和/或可以在癌症诊断之前检测到 FSM。与 LS cfDNA （1） 的发现一致，在 cfDNA 中检测到的 FSM 数量低于肿瘤中，尤其是患者 1。FSM 的概况与之前报道的 （3， 4） 和我们在 LS （1） 中观察到的相似。此外，cfDNA 中只有 1 个 FSM 的 DAF 为 >0.05 （ASTE2 C/CT;DAF = 0.076;补充表 3），表明 FSM 未在 cfDNA 中富集。单个位点的 DAF 具有特征性，序列上下文可能是一个重要的决定因素。

进一步分析显示，在患者 1 和 2 的 cfDNA 中检测到的基因中分别有 94% 和 42% 与配对的肿瘤样本共享（表 1）。cfDNA 中的低 DAF 可能部分解释了血液中检测到的 FSM 数量与肿瘤中检测到的 FSM 数量之间的不一致。此外，肿瘤异质性、患者 2 的既往肿瘤以及 CMMRD 患者从很小的时候开始所有身体细胞脱落 DNA 片段的可能性也可能是不一致的原因。

为了确定在 CMMRD 患者中检测到的 FSM 数量是否高于 LS 患者，还使用相同的平台对 3 名具有 2 对 cfDNA 的年轻 LS 患者的肿瘤（参见补充方法中的患者和标本）进行了测序（补充表 4）。LS 肿瘤和 cfDNA 中 FSM 的数量要低得多 （n = 9-33 [面板中的 7%-21% 的基因] vs. n = 8-10 [面板中的 7%-8% 的基因]，尽管 LS 肿瘤中存在杂合性的 MMR 基因丢失。这表明 FSM 可能在很小的时候从 CMMRD 患者的所有身体细胞中出现和积累，并且与 LS 患者相比，CMMRD 患者血液中可能脱落和积累的 DNA 片段更多。

治疗 CMMRD 患者具有挑战性，因为多种原发性癌症可能在生命过程中出现并影响各种不同的组织。这增加了疾病监测和管理的复杂性 （5），这突出了加强预防和监测策略的必要性。我们的研究提供了 CMMRD 中存在 FSM 的证据，并表明 FSM 不是特定于综合征或癌症类型的。需要进一步的临床前和临床评估，以确定基于移码新抗原的疫苗和为 LS 开发的生物标志物组是否对 CMMRD 有效。

 补充资料

 查看补充数据

 脚注

利益冲突：作者已声明不存在利益冲突。

版权所有：© 2024 年，Song et al.这是一篇根据 Creative Commons Attribution 4.0 International License 条款发布的开放获取文章。

提交：2024 年 8 月 14 日;接受：2024 年 10 月 15 日;发表：2024 年 10 月 22 日。

参考资料：J Clin Invest。2024;134（24）：e185999。

 引用

1. 
   Song Y 等人。外周血移码突变作为监测 Lynch 综合征的生物标志物。J Natl 癌症研究所2024;116（6）：957–965.通过以下方式查看此文章： CrossRefPubMedGoogle Scholar
2. 
   Ercan AB 等人。体质错配修复缺陷综合征的临床和生物学景观：一项国际复制修复缺陷联盟队列研究。柳叶刀 Oncol。2024;25（5）：668–682.通过以下方式查看此文章： CrossRefPubMedGoogle Scholar
3. 
   Das A 等人。种系 DNA 复制修复缺陷儿童对 PD-1 抑制反应的基因组预测因子。国家医学。2022;28（1）：125–135.通过以下方式查看此文章： CrossRefPubMedGoogle Scholar
4. 
   Bajwa-Ten Broeke SW 等人。PMS2 缺陷型结直肠癌的编码微卫星突变谱。Exp Mol Pathol.2021;122:104668.通过以下方式查看此文章： CrossRefPubMedGoogle Scholar
5. 
   DNA 聚合酶和错配修复在正常和恶性人类细胞中发挥不同的微卫星不稳定性特征。癌症 Discov。2021;11（5）：1176–1191.通过以下方式查看此文章： CrossRefPubMedGoogle Scholar