To the Editor,

Addressing the problem of heterogeneity among primary isolates of human bone marrow‐derived multipotent stromal cells (hBMSC), the recent study corresponded by Cool and Stanton et al identified a biomarker correlated with enhanced culturability.1 This represents a valuable phenotype, given hundreds of clinical trials that would benefit from early a priori knowledge that therapeutic cells would be likely to readily reach a clinical dose.2 Global transcriptome profiling of age and sex‐matched human donors linked genetic events to high and low‐growth capacity criteria; in particular, proliferative state, clonogenic potential, and telomere length. Loss of the haplotype‐specific gene encoding glutathione S‐transferase theta 1 (GSTT1), known to be absent from 38% of the population,3 was found in hBMSC of high proliferative potential, whereas cells with low‐growth capacity were GSTT1 positive.

We previously engineered hBMSC to express human telomerase (TERT),4 obtaining remarkable proliferative potential and full capacity to form heterotopic bone in murine xenograft‐scaffold models.5 Independently maintained cell strains were cultured using distinct passage split ratios, the most successful being 1:4 and 1:20, generating the hBMSC‐TERT4 and hBMSC‐TERT20 strains, respectively. Ectopic TERT expression was considered a benign means of immortalizing cells.6 Unexpectedly, after 256 population doublings during 2.5 years of continuous culture, hBMSC‐TERT20 provided the first example of spontaneous tumorigenesis in telomerised human stem cells.7 To explore aberrant events, hBMSC‐TERT20 was subcloned and extensively characterized,8 with subsequent scrutiny of any chromosome copy number variations (CNVs) that may have accumulated over time. We confirmed a microdeletion at the chromosome 9p21.3 cell cycle checkpoint gene locus, identified two potential pre‐tumorigenic biomarkers, and highlighted a network of genetic changes implicated in tumorigenic progression.9 In particular, a CNV consistently identified in hBMSC‐TERT4, pre‐and post‐tumorigenic hBMSC‐TERT20 populations and six sublones, was loss in chromosome 22q11.23 mapped to the GSTT1 gene locus (Figure 1).

Sathiyanathan et al recalled that more clonogenic hBMSC had enhanced in vivo bone forming ability and suggested that GSTT1‐null hBMSC may also be protected from the effects of oxidative stress and DNA damage. Indeed, hBMSC‐TERT populations showed better bone forming potential relative to primary cells and the hBMSC‐TERT20 cells, although compromised by discrete CNV, maintained a diploid karyotype.10 Citing ongoing debate that a GSTT1‐null phenotype may increase susceptibility to cancer, Sathiyanathan et al proposed future animal‐based studies to test this. Our data would suggest that cancer risk from GSTT1 loss is not high. hBMSC‐TERT4 cells failed to form heterotopic tumors, even when implanted with matrigel rather that osteogenic scaffold. hBMSC‐TERT20 were only tumorigenic after an unusually prolonged period of continuous culture and in addition to GSTT1 locus loss, more canonical neoplastic changes were found (eg, a microdeletion spanning three cell cycle checkpoint genes and loss of a tumor suppressor gene9). Our supportive findings lend weight to the argument that GSTT1 loss may serve as a useful genomic DNA biomarker to prospectively screen human donors before surgical isolation of bone marrow for hBMSC collection.

致编辑

为了解决人类骨髓源性多能基质细胞（hBMSC）初级分离株之间的异质性问题，Cool和Stanton等人最近进行的研究确定了与可培养性增强相关的生物标记。1这代表了一种有价值的表型，因为数百项临床试验将从早期的先验知识中受益，即治疗性细胞很容易达到临床剂量。2对年龄和性别匹配的人类供体的全球转录组谱分析将遗传事件与高和低生长能力标准联系在一起；特别是增殖状态，克隆形成潜能和端粒长度。编码谷胱甘肽S-转移酶theta 1（GSTT1）的单倍型特异性基因的丢失）（已知在38％的人口中不存在）在高增殖潜能的hBMSC中发现3个，而低生长能力的细胞则为GSTT1阳性。

我们以前曾设计过hBMSC以表达人类端粒酶（TERT），4在鼠异种移植支架模型中具有显着的增殖潜力和形成异位骨的全部能力。5使用不同的传代分裂率培养独立维持的细胞株，最成功的是1：4和1:20，分别产生了hBMSC-TERT4和hBMSC-TERT20株。异位TERT表达被认为是使细胞永生的良性手段。6出乎意料的是，在连续培养2.5年中256个种群倍增之后，hBMSC-TERT20提供了端粒化人类干细胞自发肿瘤发生的第一个例子。7为了探讨异常事件，对hBMSC‐TERT20进行了亚克隆并进行了广泛表征，参照图8，随后检查可能随时间累积的任何染色体拷贝数变异（CNV）。我们确认了染色体9p21.3细胞周期检查点基因位点的微缺失，鉴定了两个潜在的致瘤前生物标记，并强调了涉及致癌性进展的遗传变化网络。9特别是，在hBMSC-TERT4，成瘤前和后hBMSC-TERT20群体和六个亚克隆中一致鉴定出的CNV在映射到GSTT1基因位点的染色体22q11.23中丢失（图1）。

Sathiyanathan等人回忆说，更多的克隆性hBMSC增强了体内的骨形成能力，并暗示GSTT1- null hBMSC也可能受到氧化应激和DNA损伤的影响。确实，相对于原代细胞，hBMSC-TERT群体显示出更好的骨形成潜能，而hBMSC-TERT20细胞尽管受到离散CNV的损害，但仍保持二倍体核型。10 Sathiyanathan等人引用了有关GSTT1空表型可能会增加对癌症的易感性的争论，提出了未来基于动物的研究来对此进行测试。我们的数据表明，GSTT1致癌风险损失不高。即使植入基质胶而不是成骨支架，hBMSC‐TERT4细胞也无法形成异位肿瘤。hBMSC‐TERT20仅在异常延长的连续培养后才具有致癌性，除了GSTT1基因座丢失外，还发现了更多的典型肿瘤变化（例如，跨越三个细胞周期检查点基因的微缺失和肿瘤抑制基因的丢失9）。我们的支持性研究结果支持以下论点：GSTT1缺失可能是有用的基因组DNA生物标记，可在通过手术分离骨髓收集hBMSC之前对人类供体进行前瞻性筛选。