Yang Liu

← Primer

HSC · Healing Source Code

HSC 扩增的真问题——扩的是 CD34⁺,还是长期重建能力 The real problem with HSC expansion—are you expanding CD34⁺ cells, or long-term repopulating ability?

2026-07-16 · HSC扩增, 脐带血, UM171, potency, 制造

2010 年,一篇发表在《Science》上的论文里藏着两个数字。研究者往人的造血干细胞培养体系里加了一个小分子,培养结束时,带 CD34 标记的细胞多了五十倍;而这些细胞里真正还能在免疫缺陷小鼠体内落户、重建造血的那一部分,只多了十七倍。同一皿细胞,同一次培养,两把尺子给出了两个答案——五十和十七之间的那道缝隙,就是整个 HSC 扩增领域二十年来真正在处理的问题。细胞数量是可以被”养”出来的;干性不是。

为什么非扩不可

想把造血干细胞在体外养多一点,动机朴素得几乎不需要解释:很多时候,细胞根本不够用。

脐带血是这个矛盾最尖锐的地方。它随时可取、对 HLA 配型的宽容度高、慢性移植物抗宿主病的风险低,唯一的硬伤是一份脐带血里的细胞数是出生时就定死的——那么点体积,采完就没了。细胞剂量偏低带来的代价直接写在临床上:中性粒细胞恢复得慢,常常要等三周以上,而这段没有免疫防线的空窗期,正是感染与死亡集中发生的时间;移植失败的风险也更高。临床上一度靠同时输两份脐带血来凑够剂量,但这只是把问题摊开,并没有解决它。

同样的算术也压在基因治疗的制造流程上。ex vivo 的 HSC 基因治疗要先把患者自己的 CD34⁺ 细胞采出来,在体外转导或编辑,再回输。整条链上处处在损耗细胞:采集本身有上限,分选要丢一部分,培养和编辑操作又要丢一部分。如果能在体外把细胞扩出来,采集压力、放行标准、乃至整批产品的成本,都会跟着松一口气。

于是”把 HSC 养多一点”成了一个横跨脐带血移植与基因治疗制造的共同愿望。难的地方在于,它听上去像个细胞培养问题,实际上是个身份问题。

培养皿里那场沉默的流失

造血干细胞在培养里从不闲着。给它足够的细胞因子,它会分裂;但分裂之后的两个女儿细胞是不是还都是干细胞,是另一回事。培养倾向于把它推向分化:细胞总数在涨,其中真正具备 self-renewal 与长期重建能力的那一小群,却在被稀释。孔里越来越热闹,能干活的那几个越来越稀。

麻烦的是,肉眼、流式、细胞计数,都看不见这件事。CD34 是一个表型标记,不是功能证书——它标记的是一个包含干细胞在内的混杂群体,而且这层标记在培养中并不忠诚。1997 年,Bhatia 等人在《PNAS》上用免疫缺陷小鼠的重建能力反过来定义并纯化人的原始造血细胞,把 SCID-repopulating cell(SRC)确立为一个功能读数;而 2000 年,Dorrell 等人在《Blood》上给出了一个几乎是本文标题的结论:在 retroviral 转导所需的体外培养中,CD34⁺CD38⁻ 细胞的数量扩了,SRC 的频率却没有相应增加——他们把这称作 SRC 表型与功能的解离。2005 年,Zhang 与 Lodish 在《Blood》上于小鼠体系里补上了另一半:HSC 在 ex vivo 扩增过程中会改变自己的表面表型。也就是说,你手里那张用来数细胞的身份证,在培养箱里是会变造的。

这就把领域逼到了一个笨办法上:唯一算数的裁判是移植本身。把细胞按不同剂量梯度输进免疫缺陷小鼠,数月后看有没有长期、多谱系的人源造血,再用极限稀释反推出”能重建的细胞”到底有多少个。2011 年,Notta 等人在《Science》上把这套读数推到了单细胞分辨率,证明单个人 HSC 就足以支撑长期多谱系重建。这套 xenotransplant 读数严谨、也昂贵——它要花掉几个月,而这一点后面会变成整篇文章里最现实的那个死结。

两个小分子

真正的转折来自把这把功能尺子摆在筛选的出口处。

2010 年,Boitano 等人在《Science》上报告了一次以人原代 HSC 为对象的无偏筛选,捞出一个嘌呤衍生物 StemRegenin-1(SR1)。用 SR1 培养后,CD34⁺ 细胞增加约五十倍,而真能在免疫缺陷小鼠体内落户、重建造血的那一部分,只增加约十七倍——后者不是数出来的,是移植出来的:该文用极限稀释配合 Poisson 统计推算 SRC 数目,并做了二次移植验证;机制上,SR1 的作用来自拮抗 aryl hydrocarbon receptor(AhR)。这是第一次有人指出,HSC 在培养里的”掉干性”并非不可逆的宿命,而是可以被一个通路调节的过程。那五十与十七的落差,同时也诚实地记下了这条路的代价:扩出来的细胞里,大部分并不是你想要的那一群。

2014 年,Fares 等人在《Science》上报告了另一族化合物,原型叫 UM171。它扩增出的脐带血细胞在免疫缺陷小鼠里能重建人的造血至少六个月——这已经是长期重建的门槛;而更耐人寻味的是他们的一句判断:UM171 的活性独立于对 AhR 的抑制,而 AhR 那条通路作用的对象,是再生潜能更有限的那部分细胞。两个分子由此不再是同一件事的两个版本,而是分别抓住了这个混杂群体的不同层次。UM171 究竟怎么起效,直到它进入临床仍未完全讲清楚;但 2017 年 Fares 等人在《Blood》上找到了一个实用的抓手:EPCR 可以在培养中标记出经 UM171 扩增的、真正具干细胞特征的那一小群 CD34⁺ 细胞——这是在培养箱里第一次有了一张比 CD34 更靠谱的临时身份证。

走到病床边,以及终点选择的诚实

有意思的是,最早把”扩增脐带血”送进人体的,并不是这两个分子。

2010 年,Delaney 等人在《Nature Medicine》上报告了一套 Notch 配体介导的扩增体系:CD34⁺ 细胞平均扩了 164 倍,十名高危白血病患者同时接受一份未处理和一份扩增的脐带血,中性粒细胞恢复至 500/μL 的中位时间是 16 天,而同期接受双份未处理脐带血的对照队列是 26 天。加速是真实的。但这篇论文最有价值的部分恰恰是它没有回避的那一半:早期的髓系恢复几乎全部来自扩增单位,而长期看,扩增单位在多数患者体内并未留下——只有两名患者观察到较久的持续存在,其中一位到一年时也已消失。作者自己把两种解释并排摆着:可能是培养中丢了 self-renewal,也可能是未处理单位里的免疫细胞把它排掉了。

这个模式此后反复出现。2016 年,Wagner 等人在《Cell Stem Cell》上报告了 SR1 扩增脐带血的 I/II 期试验:CD34⁺ 细胞扩了 330 倍,17 名患者全部植入,中性粒细胞中位 15 天、血小板 49 天,没有一例移植失败。2020 年,Cohen 等人在《Lancet Haematology》上报告了 UM171 的 1-2 期研究,把野心推得更远——只输一份经 UM171 扩增的脐带血,22 名患者中位 18 天中性粒细胞恢复至 500/μL,无一例植入失败,由此让原本因细胞数太少而被弃用的小份脐带血重新可用。走得最远的是 omidubicel:一份脐带血在 nicotinamide 加细胞因子中培养 21 天,Horwitz 等人 2021 年在《Blood》上报告的 3 期随机试验显示,中性粒细胞植入中位 12 天,对照的未处理脐带血是 22 天,同时血小板恢复、B 细胞与 NK 细胞重建更快,病毒感染更少(10% vs 26%),细菌或侵袭性真菌感染的合并发生率也更低(37% vs 57%)。2023 年 4 月 17 日,FDA 批准了 omidubicel(商品名 Omisirge),用于计划接受脐带血移植的血液恶性肿瘤患者,以缩短中性粒细胞恢复时间、降低感染发生率——扩增脐带血由此第一次成为一款获批的产品。2025 年 12 月 8 日,FDA 又批准它用于六岁及以上、接受减低强度预处理且没有合适供者的重型再生障碍性贫血患者——扩增脐带血的适应症就此从恶性血液病扩到了骨髓衰竭。

值得停下来读一遍的,是这些试验的主要终点:中性粒细胞恢复时间、感染、植入失败。它们衡量的都是早期造血重建——也就是扩增最擅长做到的那件事。加速植入本身有确凿的临床价值,把患者从最危险的那两三周里拉出来;但把终点选在这里,同时也意味着最初那个问题被绕开了,而不是被回答了:扩出来的那些细胞,十年后还在造血吗?

效力究竟怎么量

于是问题以另一副面孔回到制造车间里:一批细胞放行之前,你怎么知道它到底有多”干”?

现实的约束近乎残酷。放行判定要在几天内做完——细胞在等着回输,患者可能已经开始清髓。而唯一被公认能测出长期重建能力的读数,是极限稀释的 xenotransplant,它要几个月。于是所有真正用于放行的指标都是替代品:CD34⁺ 细胞数、活率、扩增倍数、编辑效率、VCN。它们各自测的都是真东西,却没有一个直接测的是”能不能在人体里造血十年”。CD34 计数尤其尴尬——它恰恰是那个在培养里会变造的标记;而 SR1 那五十与十七的对照说明,倍数越漂亮,离功能反而可能越远。EPCR 这类新标记、以及单细胞层面的分子特征,正是奔着填这条缝去的,但把一个标记变成一个被监管接受的 potency assay,还需要它与人体长期重建结果之间建立起可量化的关联。(potency 这件事本身值得单独一篇,留到讲放行与效力时再展开。)

这条缝对基因治疗的意义比对脐带血更直接。2024 年,一项发表在《Blood Advances》上的工作把 UM171 用在了取自镰刀型贫血患者的、经基因修饰的造血干细胞上,报告它降低了转导培养中的 DNA 损伤、活性氧与凋亡,并在免疫缺陷小鼠中改善了这些细胞的植入与克隆多样性——注意这最后一步仍是小鼠里的临床前结果,不是人体数据。如果这条路走通,它同时松开制造流程上的两个结:采集端可以少要一些细胞,产品端可以给出更足的剂量。但也正因为基因治疗要的是一次回输、终身有效,它比任何适应症都更没有资格用”早期植入更快”来交差。这里的 potency,必须是长期的。

回到那两个数字

五十和十七之间的距离,与其说是一个技术缺口,不如说是一次提醒:细胞治疗里,数量是最容易测的量,也因此最容易被误当成目标。培养箱能可靠地给你更多细胞,却不能可靠地给你更多干细胞;而移植疗效依赖的,恰恰是后者中最不起眼、最难数清的那一小群。

过去十五年的进展是实在的:AhR 通路被找到,UM171 从筛选走进了临床,扩增脐带血成了药,基因治疗的制造开始借用同样的工具。真正还没被交上来的答卷,是一把既快得能用于放行、又准得能预言十年后骨髓里发生什么的尺子。在那把尺子出现之前,每一批被放行的细胞,某种意义上都是拿短期指标对长期结局的一次外推——这个领域最需要的下一个突破,或许不是又一个能把细胞养得更多的分子,而是一次终于能把”干性”当场称重的测量。

参考文献

  1. Bhatia M, Wang JC, Kapp U, Bonnet D, Dick JE. Purification of primitive human hematopoietic cells capable of repopulating immune-deficient mice. Proc Natl Acad Sci U S A. 1997;94(10):5320-5. DOI
  2. Dorrell C, Gan OI, Pereira DS, Hawley RG, Dick JE. Expansion of human cord blood CD34⁺CD38⁻ cells in ex vivo culture during retroviral transduction without a corresponding increase in SCID repopulating cell (SRC) frequency: dissociation of SRC phenotype and function. Blood. 2000;95(1):102-10. PubMed
  3. Zhang CC, Lodish HF. Murine hematopoietic stem cells change their surface phenotype during ex vivo expansion. Blood. 2005;105(11):4314-20. DOI
  4. Delaney C, Heimfeld S, Brashem-Stein C, Voorhies H, Manger RL, Bernstein ID. Notch-mediated expansion of human cord blood progenitor cells capable of rapid myeloid reconstitution. Nat Med. 2010;16(2):232-6. DOI
  5. Boitano AE, Wang J, Romeo R, et al. Aryl hydrocarbon receptor antagonists promote the expansion of human hematopoietic stem cells. Science. 2010;329(5997):1345-8. DOI
  6. Notta F, Doulatov S, Laurenti E, Poeppl A, Jurisica I, Dick JE. Isolation of single human hematopoietic stem cells capable of long-term multilineage engraftment. Science. 2011;333(6039):218-21. DOI
  7. Fares I, Chagraoui J, Gareau Y, et al. Cord blood expansion. Pyrimidoindole derivatives are agonists of human hematopoietic stem cell self-renewal. Science. 2014;345(6203):1509-12. DOI
  8. Wagner JE Jr, Brunstein CG, Boitano AE, et al. Phase I/II Trial of StemRegenin-1 Expanded Umbilical Cord Blood Hematopoietic Stem Cells Supports Testing as a Stand-Alone Graft. Cell Stem Cell. 2016;18(1):144-55. DOI
  9. Fares I, Chagraoui J, Lehnertz B, et al. EPCR expression marks UM171-expanded CD34⁺ cord blood stem cells. Blood. 2017;129(25):3344-3351. DOI
  10. Cohen S, Roy J, Lachance S, et al. Hematopoietic stem cell transplantation using single UM171-expanded cord blood: a single-arm, phase 1-2 safety and feasibility study. Lancet Haematol. 2020;7(2):e134-e145. DOI
  11. Horwitz ME, Stiff PJ, Cutler C, et al. Omidubicel vs standard myeloablative umbilical cord blood transplantation: results of a phase 3 randomized study. Blood. 2021;138(16):1429-1440. DOI
  12. Liu B, Klatt D, Zhou Y, Manis JP, Sauvageau G, Pellin D, Brendel C, Williams DA. UM171 enhances fitness and engraftment of gene-modified hematopoietic stem cells from patients with sickle cell disease. Blood Adv. 2024;8(22):5885-5895. DOI
  13. U.S. Food and Drug Administration. FDA approves omidubicel to reduce time to neutrophil recovery and infection in patients with hematologic malignancies. April 17, 2023. FDA
  14. U.S. Food and Drug Administration. FDA Approves First Cellular Therapy to Treat Patients with Severe Aplastic Anemia. December 8, 2025. FDA

Hidden inside a 2010 paper in Science are two numbers. The researchers added a small molecule to a culture system of human hematopoietic stem cells, and by the end of the culture, cells carrying the CD34 marker had increased fifty-fold; but the fraction of those cells that could still home to an immunodeficient mouse and rebuild its blood had increased only seventeen-fold. Same dish of cells, same culture, two rulers giving two different answers—and the gap between fifty and seventeen is exactly the problem the entire field of HSC expansion has spent twenty years working on. Cell number is something you can “culture up”; stemness is not.

Why expansion is unavoidable

The motive for growing more hematopoietic stem cells outside the body is so plain it barely needs explaining: much of the time, there simply aren’t enough cells.

Cord blood is where this tension is sharpest. It is available on demand, it is forgiving about HLA matching, and it carries a low risk of chronic graft-versus-host disease; its one hard limitation is that the number of cells in a single cord blood unit was fixed at birth—that small a volume, and once it’s collected, that’s all there is. The price of a low cell dose is written directly into the clinic: neutrophils recover slowly, often taking more than three weeks, and that window without an immune defense is precisely when infection and death cluster; the risk of graft failure is higher too. For a time the clinic resorted to infusing two cord blood units at once to reach an adequate dose, but that only spreads the problem out—it does not solve it.

The same arithmetic bears down on the manufacturing process for gene therapy. Ex vivo HSC gene therapy first has to collect the patient’s own CD34⁺ cells, transduce or edit them outside the body, and then reinfuse them. Cells are lost at every point along that chain: collection itself has a ceiling, sorting discards a portion, and the culture and editing steps discard another. If the cells could be expanded ex vivo, the pressure on collection, the release criteria, and even the cost of an entire batch of product would all get some breathing room.

So “grow more HSCs” became a shared wish spanning cord blood transplantation and gene therapy manufacturing. The difficulty is that it sounds like a cell culture problem when it is actually a problem of identity.

The silent attrition in the dish

Hematopoietic stem cells are never idle in culture. Give them enough cytokines and they will divide; but whether both daughter cells after that division are still stem cells is another matter. Culture tends to push them toward differentiation: the total cell count climbs, while the small population that genuinely possesses self-renewal and long-term repopulating ability is diluted out. The well gets more and more crowded, and the few that can do the work get rarer and rarer.

The trouble is that the naked eye, flow cytometry, and cell counting all fail to see this. CD34 is a phenotypic marker, not a certificate of function—it marks a mixed population that contains stem cells among other things, and that marker is not faithful in culture. In 1997, Bhatia and colleagues, writing in PNAS, used repopulating ability in immunodeficient mice to define and purify primitive human hematopoietic cells in reverse, establishing the SCID-repopulating cell (SRC) as a functional readout; and in 2000, Dorrell and colleagues, in Blood, delivered a conclusion that is almost the title of this piece: in the ex vivo culture required for retroviral transduction, the number of CD34⁺CD38⁻ cells expanded, but SRC frequency did not increase correspondingly—they called this a dissociation of SRC phenotype and function. In 2005, Zhang and Lodish supplied the other half in Blood, in the mouse system: HSCs change their own surface phenotype during ex vivo expansion. In other words, the ID card you are holding to count cells can be forged inside the incubator.

This forced the field onto a clumsy method: the only judge that counts is transplantation itself. Infuse the cells into immunodeficient mice across a range of doses, look months later for long-term, multilineage human hematopoiesis, and then use limiting dilution to work backward to how many “cells capable of repopulating” there actually were. In 2011, Notta and colleagues, in Science, pushed this readout to single-cell resolution, proving that a single human HSC suffices to support long-term multilineage reconstitution. This xenotransplant readout is rigorous, and it is expensive—it takes months, a fact that later becomes the most concrete deadlock in this whole piece.

Two small molecules

The real turning point came from putting that functional ruler at the exit of the screen.

In 2010, Boitano and colleagues reported in Science an unbiased screen carried out on primary human HSCs, which pulled out a purine derivative, StemRegenin-1 (SR1). After culture with SR1, CD34⁺ cells increased roughly fifty-fold, while the fraction that could genuinely home to and rebuild blood in immunodeficient mice increased only roughly seventeen-fold—and the latter was not counted, it was transplanted: the paper used limiting dilution together with Poisson statistics to estimate SRC numbers, and performed secondary transplantation for validation; mechanistically, SR1 acts by antagonizing the aryl hydrocarbon receptor (AhR). This was the first time anyone had pointed out that “losing stemness” in culture is not an irreversible fate for HSCs but a process that can be modulated through a pathway. That gap between fifty and seventeen also honestly records the cost of this route: most of the cells you expand are not the population you wanted.

In 2014, Fares and colleagues reported in Science another family of compounds, the prototype called UM171. The cord blood cells it expands can reconstitute human hematopoiesis in immunodeficient mice for at least six months—already the threshold for long-term reconstitution; and more intriguing is a judgment of theirs: UM171’s activity is independent of AhR inhibition, and the cells the AhR pathway acts on are the ones with more limited regenerative potential. The two molecules thereby stopped being two versions of the same thing and became distinct grips on different layers of this mixed population. Exactly how UM171 works was still not fully spelled out even as it entered the clinic; but in 2017 Fares and colleagues, in Blood, found a practical handle: EPCR can mark, in culture, the small subset of UM171-expanded CD34⁺ cells that genuinely carry stem cell characteristics—the first time the incubator had an interim ID card more reliable than CD34.

To the bedside, and the honesty of endpoint selection

Interestingly, the first to take “expanded cord blood” into humans was neither of these two molecules.

In 2010, Delaney and colleagues reported in Nature Medicine a Notch ligand-mediated expansion system: CD34⁺ cells expanded an average of 164-fold, and ten patients with high-risk leukemia each received one unmanipulated and one expanded cord blood unit simultaneously; the median time to neutrophil recovery to 500/μL was 16 days, versus 26 days in a contemporaneous control cohort receiving two unmanipulated cord blood units. The acceleration is real. But the most valuable part of this paper is precisely the half it did not dodge: early myeloid recovery came almost entirely from the expanded unit, while over the long run the expanded unit did not persist in most patients—longer persistence was observed in only two patients, and in one of them it had disappeared by one year. The authors laid two explanations side by side: self-renewal may have been lost in culture, or immune cells in the unmanipulated unit may have rejected it.

This pattern recurred repeatedly thereafter. In 2016, Wagner and colleagues reported in Cell Stem Cell a phase I/II trial of SR1-expanded cord blood: CD34⁺ cells expanded 330-fold, all 17 patients engrafted, median neutrophil recovery was 15 days and platelets 49 days, with no case of graft failure. In 2020, Cohen and colleagues reported in Lancet Haematology a phase 1-2 study of UM171 that pushed the ambition further—infusing only a single UM171-expanded cord blood unit, 22 patients reached neutrophil recovery to 500/μL at a median of 18 days with no case of graft failure, thereby making usable again the small cord blood units that had been discarded for having too few cells. The one that went furthest is omidubicel: a cord blood unit cultured for 21 days in nicotinamide plus cytokines. The phase 3 randomized trial reported by Horwitz and colleagues in Blood in 2021 showed median neutrophil engraftment at 12 days versus 22 days for unmanipulated cord blood as the control, along with faster platelet recovery and faster B cell and NK cell reconstitution, fewer viral infections (10% vs 26%), and a lower combined incidence of bacterial or invasive fungal infections (37% vs 57%). On April 17, 2023, the FDA approved omidubicel (brand name Omisirge) for patients with hematologic malignancies planned for cord blood transplantation, to shorten time to neutrophil recovery and reduce the incidence of infection—making expanded cord blood an approved product for the first time. On December 8, 2025, the FDA further approved it for patients aged six and older with severe aplastic anemia who receive reduced-intensity conditioning and have no suitable donor—extending the indication for expanded cord blood from hematologic malignancy to bone marrow failure.

What is worth pausing to read is the primary endpoints of these trials: time to neutrophil recovery, infection, graft failure. All of them measure early hematopoietic reconstitution—which is exactly what expansion is best at delivering. Accelerated engraftment has definite clinical value in its own right, pulling patients out of those most dangerous two or three weeks; but choosing the endpoint here also means the original question was circumvented rather than answered: are those expanded cells still making blood ten years later?

How, exactly, do you measure potency

So the question comes back to the manufacturing floor wearing another face: before a batch of cells is released, how do you know how “stem” it really is?

The real constraints are close to brutal. The release decision has to be made within days—the cells are waiting to be reinfused, and the patient may already have begun myeloablation. Yet the one readout acknowledged to measure long-term repopulating ability is the limiting-dilution xenotransplant, and it takes months. So every metric actually used for release is a surrogate: CD34⁺ cell number, viability, fold expansion, editing efficiency, VCN. Each of them measures something real, yet not one of them directly measures “can this make blood in a human being for ten years.” The CD34 count is especially awkward—it is precisely the marker that gets forged in culture; and the fifty-versus-seventeen contrast from SR1 shows that the prettier the fold expansion, the further it may be from function. New markers such as EPCR, along with molecular signatures at the single-cell level, are aimed squarely at filling this gap, but turning a marker into a potency assay accepted by regulators still requires establishing a quantifiable relationship between it and long-term repopulating outcomes in humans. (Potency is itself worth a piece of its own; that will wait until we take up release and potency.)

This gap matters more directly for gene therapy than for cord blood. In 2024, a study published in Blood Advances applied UM171 to gene-modified hematopoietic stem cells taken from patients with sickle cell disease, reporting that it reduced DNA damage, reactive oxygen species, and apoptosis during transduction culture, and in immunodeficient mice improved the engraftment and clonal diversity of these cells—note that this last step is still a preclinical result in mice, not human data. If this route works out, it loosens two knots in the manufacturing process at once: the collection end can ask for fewer cells, and the product end can deliver a fuller dose. But precisely because what gene therapy wants is one infusion, lifelong effect, it is less entitled than any other indication to settle up with “faster early engraftment.” Here, potency has to be long-term.

Back to those two numbers

The distance between fifty and seventeen is less a technical gap than a reminder: in cell therapy, quantity is the easiest thing to measure, and therefore the easiest thing to mistake for the goal. An incubator can reliably give you more cells, but it cannot reliably give you more stem cells; and what transplant efficacy depends on is precisely the most inconspicuous, hardest-to-count small population among the latter.

The progress of the past fifteen years is real: the AhR pathway was found, UM171 went from a screen into the clinic, expanded cord blood became a drug, and gene therapy manufacturing began borrowing the same tools. What still has not been handed in is a ruler both fast enough to be used for release and accurate enough to predict what will happen in the bone marrow ten years from now. Until that ruler exists, every batch of cells that gets released is, in a sense, an extrapolation from short-term metrics to a long-term outcome—and the next breakthrough this field most needs may not be yet another molecule that grows more cells, but a measurement that can finally weigh “stemness” on the spot.

References

  1. Bhatia M, Wang JC, Kapp U, Bonnet D, Dick JE. Purification of primitive human hematopoietic cells capable of repopulating immune-deficient mice. Proc Natl Acad Sci U S A. 1997;94(10):5320-5. DOI
  2. Dorrell C, Gan OI, Pereira DS, Hawley RG, Dick JE. Expansion of human cord blood CD34⁺CD38⁻ cells in ex vivo culture during retroviral transduction without a corresponding increase in SCID repopulating cell (SRC) frequency: dissociation of SRC phenotype and function. Blood. 2000;95(1):102-10. PubMed
  3. Zhang CC, Lodish HF. Murine hematopoietic stem cells change their surface phenotype during ex vivo expansion. Blood. 2005;105(11):4314-20. DOI
  4. Delaney C, Heimfeld S, Brashem-Stein C, Voorhies H, Manger RL, Bernstein ID. Notch-mediated expansion of human cord blood progenitor cells capable of rapid myeloid reconstitution. Nat Med. 2010;16(2):232-6. DOI
  5. Boitano AE, Wang J, Romeo R, et al. Aryl hydrocarbon receptor antagonists promote the expansion of human hematopoietic stem cells. Science. 2010;329(5997):1345-8. DOI
  6. Notta F, Doulatov S, Laurenti E, Poeppl A, Jurisica I, Dick JE. Isolation of single human hematopoietic stem cells capable of long-term multilineage engraftment. Science. 2011;333(6039):218-21. DOI
  7. Fares I, Chagraoui J, Gareau Y, et al. Cord blood expansion. Pyrimidoindole derivatives are agonists of human hematopoietic stem cell self-renewal. Science. 2014;345(6203):1509-12. DOI
  8. Wagner JE Jr, Brunstein CG, Boitano AE, et al. Phase I/II Trial of StemRegenin-1 Expanded Umbilical Cord Blood Hematopoietic Stem Cells Supports Testing as a Stand-Alone Graft. Cell Stem Cell. 2016;18(1):144-55. DOI
  9. Fares I, Chagraoui J, Lehnertz B, et al. EPCR expression marks UM171-expanded CD34⁺ cord blood stem cells. Blood. 2017;129(25):3344-3351. DOI
  10. Cohen S, Roy J, Lachance S, et al. Hematopoietic stem cell transplantation using single UM171-expanded cord blood: a single-arm, phase 1-2 safety and feasibility study. Lancet Haematol. 2020;7(2):e134-e145. DOI
  11. Horwitz ME, Stiff PJ, Cutler C, et al. Omidubicel vs standard myeloablative umbilical cord blood transplantation: results of a phase 3 randomized study. Blood. 2021;138(16):1429-1440. DOI
  12. Liu B, Klatt D, Zhou Y, Manis JP, Sauvageau G, Pellin D, Brendel C, Williams DA. UM171 enhances fitness and engraftment of gene-modified hematopoietic stem cells from patients with sickle cell disease. Blood Adv. 2024;8(22):5885-5895. DOI
  13. U.S. Food and Drug Administration. FDA approves omidubicel to reduce time to neutrophil recovery and infection in patients with hematologic malignancies. April 17, 2023. FDA
  14. U.S. Food and Drug Administration. FDA Approves First Cellular Therapy to Treat Patients with Severe Aplastic Anemia. December 8, 2025. FDA