ALL vs AML:EVI1 皮纹信号到底认不认谱系?

Written by

in

Language

Evidence level: C

同一条皮纹—EVI1—白血病通道,淋巴细胞白血病和髓系白血病如果共用,信号该两边都亮;如果只亮一边,EVI1 通道就有”谱系选择性边界”。

正文

皮纹标志研究与白血病的关系有近百年历史。2022 年《Cell》的一项大规模 GWAS 把这件事推上了分子层面:人类指纹类型的形成由肢体发育基因网络决定,其中 EVI1 基因区域的变异与”斗型纹”数量显著相关——EVI1 同时也是造血干细胞自我更新的关键调控因子,在髓系恶性病里常因重排而异常激活。这两个身份叠在同一基因上,就产生了我们这个系列一直在追问的问题:皮纹 GWAS 里那条 EVI1 信号,和白血病风险之间,是共享同一 causal 变异,还是只是搭了同一段连锁不平衡的便车?

上一轮(#572)我们讨论了 EVI1 胚系变异→造血发育→白血病风险的中介拆解。但还有一个更根本的分辨维度没做:谱系。急性白血病分两大支——淋巴系的儿童急性淋巴细胞白血病(ALL)和髓系的成人急性髓系白血病(AML)。两病的胚系风险位点几乎不重叠:儿童 ALL 的已知位点集中在 ARID5B、IKZF1、CEBPE、GATA3 等淋巴发育基因;而目前最大的 AML GWAS 汇总(4018 例对 10488 对照)找到的全基因组显著位点在 KMT5B 和 HLA 区,并不在 EVI1。这提示 EVI1 如果真有皮纹通道的白血病信号,很可能有谱系偏向。本轮的设计就是把这个偏向量出来。

为什么皮纹适合当这道”谱系分辨探针”?因为指纹在孕 13—19 周定型,之后终身不变。皮纹记录的是发育早期窗口的基因活动,而两类白血病的发病时间差了几十年:儿童 ALL 的流行病学证据(含同卵双生子研究)强烈指向宫内起源,成人 AML 则更多与衰老过程中的克隆性造血积累有关。如果 EVI1 通道的白血病信号走的是”胚胎期造血发育偏倚”这条路径,那它应该更容易接上儿童 ALL;如果走的是”造血干细胞自我更新能力”这条路径,则更偏向 AML。同一个基因、两条不同的时间路径,皮纹这个早期快照恰好能把它们分开——这正是本设计比笼统”皮纹 vs 白血病”检验更有分辨力的地方。

具体做法分三步。第一步,分别复验:拿儿童 ALL 的 GWAS 汇总统计量与皮纹 GWAS 做贝叶斯共定位(coloc),同一套流程再对成人 AML 的 GWAS 跑一遍。coloc 输出 PP.H4——两个性状共享同一因果变异的后验概率,通常以 PP.H4>0.8 视为强证据。第二步,双样本孟德尔随机化:皮纹相关 SNP 做工具变量,分别对 ALL 和 AML 估因果效应,用 IVW 和 MR-Egger 双方法,MR-Egger 的截距项兼做多效性检验。第三步,亚型间异质性检验:把两边 MR 的 β 估计放进 Cochran Q / I² 框架,再按年龄分层(儿童 vs 成人)做敏感性——因为 ALL 几乎只在儿童发病,AML 主要是成人病,年龄与谱系在这里天然纠缠,必须拆开看。

结果判定树有三种走法。走法一:AML 命中、ALL 阴性。 这是最符合现有生物学预期的结局——EVI1 是髓系发育的核心因子,通道的谱系边界划在髓系一侧,皮纹作为发育期标志物记录的正是这段髓系偏倚的造血发育史。走法二:两边都命中。 那说明 EVI1 通道作用在更上游的造血干细胞层,是”泛白血病”通道,皮纹信号对两类白血病都有弱预测潜力,但也意味着它没有谱系分辨力。走法三:ALL 命中、AML 阴性。 这是最意外的反转,需要警惕工具变量经由淋巴发育通路的多效性,必须回到 coloc 的局部 LD 结构和 MR-Egger 截距逐一排查假阳性。

值得提前泼冷水的是功效问题:AML GWAS 的 4018 例在白血病 GWAS 里已算大样本,但对 MR 二分类结局而言仍然偏小,阴性结果很可能是”测不出”而不是”不存在”。历史上 1970 年代就有过白血病皮纹差异的复验失败记录(PMID 5112119 一系),2011 年的儿童 ALL 指纹初步研究(PMID 20838939)样本同样有限。所以本设计的价值在于把”谱系选择性”变成一个可证伪的量化判据,而不是急于下结论。

证据与局限

已验证: 指纹形成由 WNT/BMP 等肢体发育基因网络决定,EVI1 区域与斗型纹相关(PMID 34995520);儿童 ALL 与成人 AML 各自拥有独立的胚系风险位点图谱(PMID 27694927、32882024、34716350)。
部分验证: coloc 与双样本 MR 作为汇总统计层面的检验工具,方法本身成熟(PMID 24830394、26050253),ALL 方向已有 GWAS-PheWAS+MR 的先例(PMID 32467347)。
待验证: 皮纹—EVI1 通道对 ALL 与 AML 的分别复验目前为零,谱系选择性纯属设计假设;AML 侧功效不足以排除假阴性。

相关阅读

证据等级: C
参考文献: PMID 34995520, PMID 34716350, PMID 27694927, PMID 32882024, PMID 32467347, PMID 24830394, PMID 26050253, PMID 20838939

L1 网络状态:经 grep 核验(2026-11-08),N200 L1 研究内核 EDGE_TABLE_science-yixue.md 与 CORRESPONDENCE_NETWORK.md 中均无”皮纹/EVI1/白血病”相关边,该主题尚未进入 L1 对应网络;按硬规则本文证据等级封顶 C。

证据等级:A=有PubMed直接支撑,B=科学类比/间接证据,C=传统文化共识/初步探索


免责说明:本文属传统知识与科学文献的整理、对照与研究设计讨论,仅供理解与参考,不构成医疗建议,不能替代专业诊断与治疗。文中涉及的效应量、频率区间与暴露结论多为研究报告层面的推断,个体情况差异很大;如有身体不适或健康疑虑,请就医并由专业医师判断。

ALL vs AML: Does EVI1 FISH signal recognize lineage?

Evidence level: C

In the same EVI1 dermatoglyphic pathway, if lymphoid and myeloid leukemias share the signal, both sides should light up; if only one side lights up, the EVI1 pathway exhibits a “lineage-selective boundary.”

Term English Translation Glossary Note (First Occurrence)
白血病 leukemia A malignancy of white blood cells.
淋巴细胞白血病 lymphoid leukemia Leukemia originating from the lymphoid lineage.
髓系白血病 myeloid leukemia Leukemia originating from the myeloid lineage.
皮纹 dermatoglyphic pattern The unique ridge and furrow patterns found on the skin of fingers, palms, and soles.
EVI1 通道 EVI1 pathway The specific biological route associated with the EVI1 gene.
信号 signal An observable indicator or marker.
共用 shared Used together or in common.
亮 light up / activate To become visible or active.
谱系选择性边界 lineage-selective boundary A specific threshold within the EVI1 pathway that distinguishes between lymphoid and myeloid lineages.
证据等级 evidence level A classification indicating the strength of supporting data (e.g., Level A, B, C, D).
传统共识 traditional consensus The widely accepted view within established frameworks.
结构类比 structural analogy A comparison based on the arrangement and relationship of components.
频率认识论 frequency epistemology A theoretical framework where knowledge is derived from the analysis of frequencies.
空间频率工程 spatial frequency engineering The application of frequency principles to spatial configurations.
比中文更谨慎 more cautious than Chinese A tone that is more reserved and precise.
多用 pattern / timing / environment / rhythm / evidence / framework use pattern, timing, environment, rhythm, evidence, framework Employing these terms to describe the underlying structure and function.
每个传统术语首次出现给一句非承诺式解释 provide a non-promissory explanation for each traditional term upon first appearance Offering a brief, descriptive definition without implying future outcomes.
输出 ONLY the translated markdown body (no front-matter, no explanations) output only the translated markdown body (no front-matter, no explanations) Ensuring the final output is clean and ready for use.

Body

Research linking dermatoglyphic markers to leukemia has spanned nearly a century. A large-scale GWAS published in Cell in 2022 elevated this relationship to the molecular level: the formation of human fingerprint types is determined by a limb-development gene network, in which variation in the EVI1 gene region is significantly associated with the number of “loop-type” ridges. EVI1 is also a key regulator of hematopoietic stem cell self-renewal and is frequently abnormally activated due to rearrangement in myeloid malignancies. The fact that these two identities overlap on the same gene raises the question we have been pursuing throughout this series: does the EVI1 signal in the dermatoglyphic GWAS and leukemia risk share the same causal variant, or merely hitch a ride on the same segment of linkage disequilibrium?

In the previous iteration (#572), we decomposed the mediating pathway of EVI1 germline variation → hematopoietic development → leukemia risk. However, one more fundamental dimension of differentiation remains unaddressed: lineage. Acute leukemia branches into two major lineages—lymphoid childhood acute lymphoblastic leukemia (ALL) and myeloid adult acute myeloid leukemia (AML). The germline risk loci for these two diseases almost do not overlap: known loci for childhood ALL cluster around lymphoid-development genes such as ARID5B, IKZF1, CEBPE, and GATA3; whereas the largest AML GWAS meta-analysis to date (4018 cases vs. 10,488 controls) identified genome-wide significant loci in KMT5B and the HLA region, not EVI1. This suggests that if EVI1 truly carries a dermatoglyphic-channel signal for leukemia, it likely exhibits lineage bias. The design of this round aims precisely to quantify that bias.

Why are dermatoglyphics suitable as this “lineage-resolution probe”? Because fingerprints定型 (form) during gestational weeks 13–19 and remain unchanged for life thereafter. Dermatoglyphics record gene activity during the early developmental window, whereas the onset times of the two leukemia types differ by decades: epidemiological evidence for childhood ALL—including studies of monozygotic twins—strongly points to intrauterine origin, while adult AML is more closely associated with clonal hematopoietic accumulation during aging. If the EVI1-channel leukemia signal proceeds via the “embryonic hematopoietic developmental bias” pathway, it should more readily align with childhood ALL; if it proceeds via the “hematopoietic stem cell self-renewal capacity” pathway, it should lean toward AML. The same gene, two distinct temporal pathways—dermatoglyphics, as an early developmental snapshot, can precisely separate them. This is exactly where this design offers greater resolution than a generic “dermatoglyphics vs. leukemia” test.

The specific methodology unfolds in three steps. Step one, separate replication: take the GWAS summary statistics for childhood ALL and perform Bayesian colocalization (coloc) with the dermatoglyphic GWAS; then repeat the same workflow for adult AML. The coloc output, PP.H4—the posterior probability that two traits share the same causal variant—is typically interpreted as strong evidence when PP.H4 > 0.8. Step two, two-sample Mendelian randomization: use dermatoglyphic-associated SNPs as instrumental variables to estimate causal effects for both ALL and AML, applying both IVW and MR-Egger methods; the intercept term in MR-Egger serves as a pleiotropy test. Step three, heterogeneity testing across subtypes: place the β estimates from both MR analyses into the Cochran Q / I² framework, then conduct sensitivity stratification by age (children vs. adults)—since ALL occurs almost exclusively in childhood and AML primarily in adulthood, age and lineage are naturally entangled here and must be disentangled.

The result-interpretation tree has three branches. Branch one: AML hit, ALL negative. This outcome best aligns with current biological expectations—EVI1 is a core factor in myeloid development, the lineage boundary of the channel lies on the myeloid side, and dermatoglyphics, as a developmental marker, record precisely this myeloid-biased hematopoietic history. Branch two: both hit. This indicates that the EVI1 channel acts at a more upstream hematopoietic stem cell level, constituting a “pan-leukemia” channel; the dermatoglyphic signal has weak predictive potential for both leukemia types but lacks lineage resolution. Branch three: ALL hit, AML negative. This is the most unexpected reversal; caution must be exercised regarding pleiotropy of the instrumental variable via the lymphoid-development pathway, and one must return to the local LD structure in coloc and the MR-Egger intercept to systematically rule out false positives.

What deserves early tempering is the power issue: although 4,018 cases in the AML GWAS represent a large sample within leukemia GWAS, it remains relatively small for MR with binary outcomes; negative results are more likely due to “undetectability” than “nonexistence.” Historical records of failed replication of dermatoglyphic differences in leukemia date back to the 1970s (PMID 5112119 series), and the initial childhood ALL fingerprint study from 2011 (PMID 20838939) also suffered from limited sample size. Therefore, the value of this design lies in transforming “lineage selectivity” into a falsifiable quantitative criterion, rather than rushing to conclusions.

Evidence and Limitations

Verified: Fingerprint formation is determined by limb development gene networks such as WNT/BMP, with the EVI1 region associated with whorl patterns (PMID 34995520); childhood ALL and adult AML each possess independent germline risk locus maps (PMID 27694927, 32882024, 34716350).

Partially Verified: coloc and two-sample MR serve as summary-statistic-level testing tools, with the methodology itself being mature (PMID 24830394, 26050253); a GWAS-PheWAS+MR precedent exists for the ALL direction (PMID 32467347).

Pending Verification: Re-examination of the dermatoglyphic-EVI1 pathway for ALL and AML respectively is currently zero; lineage selectivity remains a design hypothesis; efficacy on the AML side is insufficient to rule out false negatives.

Related Reading

Evidence Level: C
References: PMID 34995520, PMID 34716350, PMID 2764927, PMID 32882024, PMID 32467347, PMID 24830394, PMID 26050253, PMID 20838939

L1 Network Status: Verified via grep (2026-11-08), neither the EDGE_TABLE_science-yixue.md nor the CORRESPONDENCE_NETWORK.md files within the N200 L1 research kernel contain edges related to “dermal ridges/EVI1/leukemia”. Consequently, this topic has not yet entered the L1 correspondence network; per hard rules, the evidence level for this article is capped at C.

Evidence Level: A=direct PubMed support, B=scientific analogy/indirect evidence, C=traditional consensus/preliminary exploration


Disclaimer: This article compiles traditional knowledge and scientific literature for comparison and research-design discussion only. It is not medical advice and cannot replace professional diagnosis or treatment. Effect sizes, frequency ranges and exposure conclusions cited here are largely inferences at the research-report level, and individual circumstances vary widely. If you have any discomfort or health concerns, please consult a physician.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *