Yang Liu

← Primer

HSC · Healing Source Code

CD34 不是 HSC 的身份证——临床富集标志、异质细胞群与真正的长期重建细胞 CD34 is not the ID card of a hematopoietic stem cell — a clinical enrichment marker, a heterogeneous population, and the cells that truly reconstitute long-term

2026-10-09 · 造血干细胞, CD34, 长期重建, 异质性, 基因治疗, 表型与功能

在自体造血干细胞移植和基因编辑疗法的起始材料描述里，「CD34⁺ 细胞」几乎是个绕不开的词。久而久之，这个计数被读成了「干细胞有多少个」。可 CD34 不是一块刻着「我是造血干细胞」的身份证，它更像一件在骨髓里被广泛穿着的制服：穿它的人里，有撑起终身造血的极少数，有大批只管眼前几周的中间角色，偶尔还有人换了便装，制服不在身上，却照样担得起整个血液系统。

一个好用的把手

故事从一株抗体开始。1984 年，Civin 等人在 J Immunol 上报告了抗体 My-10：它以人的未成熟髓系细胞系 KG-1a 为免疫原制备，选择条件是不与成熟粒细胞反应，结果发现它特异地结合在未成熟的正常人骨髓细胞上，其中包括造血祖细胞，作者因此指出，这株抗体可以用来鉴定和纯化造血祖细胞。1992 年，Simmons 等人在同一本杂志上从 KG-1 细胞文库里克隆了 CD34 的 cDNA，描述它是一个约 115 kDa 的跨膜糖蛋白，功能不明，除了出现在造血祖细胞上，也出现在多种组织的小血管内皮上。

真正让这个抗原从「标记」变成「工具」的，是移植实验。1988 年，Berenson 等人在 J Clin Invest 上报告：CD34 存在于 1–4% 的人骨髓细胞上，几乎涵盖体外实验能检出的所有造血祖细胞。他们用与之反应的抗体 12-8 富集了五只狒狒的自体骨髓，富集后的 15–27 × 10⁶ 个细胞里有 65–91% 表达 CD34；动物经致死量照射后回输这些细胞，五只都在 13–24 天内让粒细胞计数超过 1,000/mm³、血小板超过 20 × 10³/mm³，随后血象恢复正常。作为对照，另外两只动物收到的是去除了 CD34⁺ 细胞的骨髓，每公斤体重收到 184 和 285 × 10⁶ 个细胞（前一组为 3.2–4.4 × 10⁶），一只在第 29 天死亡而未植活，另一只全血细胞减少超过 100 天。作者的结论措辞很克制：这些数据提示，负责造血重建的干细胞是 CD34⁺ 的。

对一个想把干细胞从骨髓里分出来用的临床领域，这已经足够。一个抗体能抓住它、一个计数能量化它、一次移植能验证它，CD34 因此成了流程里的标准动作。需要同时记住的是，它是一个把手，不是干细胞性本身的零件：1996 年，Cheng 等人在 Blood 上报告，敲除 CD34 的小鼠发育正常，成年血象看上去与野生型无异，亚致死照射后红系、髓系和血小板的恢复动力学与野生型无法区分，尽管骨髓和脾脏里祖细胞的集落形成活性明显下降；作者也指出，CD34 在祖细胞的形成中起重要作用，只是成年小鼠体内可能还留有足够的祖细胞储备。至少在成年稳态和亚致死照射后的恢复这两项上，缺了它的小鼠看不出血象差异。

牌子下面是一个分层的人群

把手好用，不等于它抓住的东西是均质的。从 1990 年代起，每一次对人造血干细胞的提纯，几乎都是在 CD34⁺ 的内部做减法。1992 年，Baum 等人在 PNAS 上从人胎儿骨髓里圈出表达 Thy-1 和 CD34、且不带已知谱系标记的细胞，它们只占 0.05–0.1%，却集中了能在小鼠基质细胞上建立长期多谱系培养的活性。2007 年，Majeti 等人在 Cell Stem Cell 上把脐带血和骨髓里的 Lin⁻CD34⁺CD38⁻ 部分再按 CD90 与 CD45RA 拆成三群：脐带血里 CD90⁺CD45RA⁻ 的那一群含有 HSC，移植实验显示活性可以收缩到只有 10 个纯化细胞的量级；CD90⁻CD45RA⁻ 的一群则首次被前瞻性地分离出来，作为候选的多能祖细胞（multipotent progenitor, MPP）。2011 年，Notta 等人在 Science 上进一步用 CD49f 区分 HSC 与只能短暂植入的 MPP：单个 CD49f⁺ 细胞就能高效产生长期多谱系移植物，失去 CD49f 的细胞则只能短暂植入。到 2022 年，Anjos-Afonso 等人在 Nat Commun 上报告了 CD34⁺EPCR⁺（CD38/CD45RA）⁻ 的群体，在异种移植实验中其干细胞频率约为每 3 个细胞中有 1 个，作者称这是迄今所描述的最高值。

从骨髓细胞的 1–4% 一路削到「三个里有一个」，说明的是同一件事：CD34⁺ 是一个入口，不是终点。而这个入口里的结构，也未必是一排整齐的台阶。2019 年，Pellin 等人在 Nat Commun 上用单细胞 RNA-seq 检验了现有的分选方案，发现按这些方案切开的 CD34⁺ 骨髓细胞确实对应一个有层级的转录景观，但这套景观会漏掉重要的早期命运决定，他们同时提出 CD164 可作为标记最早分支的候选。2017 年，Velten 等人在 Nat Cell Biol 上进一步报告：在稳态下，个体 HSC 是逐渐获得谱系倾向的，谱系限制型细胞从一个「低度倾向的未分化连续体」里直接长出来，而不是依次穿过离散的祖细胞台阶。2016 年，Notta 等人在 Science 上从单个 CD34⁺ 细胞出发，对不同发育阶段的造血做了命运图谱：胎儿肝里有大量混合髓系、红系、巨核命运的寡能祖细胞，成人骨髓里却主要剩下 multipotent 和 unipotent 两类。同一个标志之下，装的东西随发育阶段而变。造血树究竟是离散的层级还是连续的景观，是另一篇要专门拆的争论，这里只需记住一点：CD34⁺ 是一个被各种分选方案反复切分的混合群体。

没戴牌子的细胞

更棘手的是反方向的例子。1996 年，Osawa 等人在 Science 上发现，与人不同，成年小鼠骨髓里原始的 HSC 活性出现在 CD34 低表达到阴性的部分；注射单个 CD34 低或阴性、c-Kit⁺、Sca-1⁺、谱系标记阴性的细胞，在 21% 的受体里实现了淋巴造血系统的长期重建。1998 年，Bhatia 等人在 Nat Med 上报告了人源细胞里的类似现象：在 NOD/SCID 小鼠里，存在一类能起始多谱系造血的人源细胞，他们称为 CD34neg-SRC，它们局限在 Lin⁻CD34⁻CD38⁻ 的群体里，体外集落形成能力很低，并在某些短期悬浮培养条件下数量增加，而同样的条件无法维持源自 CD34⁺ 群体的 SRC。2013 年，Anjos-Afonso 等人在 Cell Stem Cell 上又把 Lin⁻CD34⁻CD38⁻CD93hi 的群体拿来比较，认为其中含有 HSC，并能把它们放在 CD34⁺ 群体之上，处于相当原始且静息的状态，依赖 Notch 等信号维持。

这些细胞都很稀少。2023 年，同一个实验室的 Anjos-Afonso 与 Bonnet 在 Blood 的综述里提到极稀有的 CD34⁻ HSC 群体被认为位于 CD34⁺ HSC 之上，同时写道，人的造血主要由后者维持。所以更准确的说法不是「CD34⁻ 才是真正的干细胞」，而是 CD34 这块牌子既会放进不是干细胞的细胞，偶尔也会把是干细胞的挡在门外。

对 CD34 在小鼠里的变化，还有一种解读，更接近「状态」而不是「身份」。1999 年，Sato 等人在 Blood 上用小鼠移植模型报告：正常成年小鼠里，具有长期植入能力的干细胞多数是 CD34⁻ 的；用 5-FU 处理后，干细胞在 CD34⁻ 和 CD34⁺ 两个部分都能找到；把谱系阴性、c-Kit⁺、Sca-1⁺、CD34⁻ 的骨髓细胞在 IL-11 和 steel factor 里体外扩增一周后，只有 CD34⁺ 的部分还能在 2 到 5 个月里植入；而把 5-FU 后的 CD34⁺ 细胞移植进去，等骨髓回到稳态，多数具有长期植入能力的细胞又回到了 CD34⁻ 的部分。作者因此认为，CD34 的表达反映的是干细胞的激活状态，并且是可逆的。2000 年，Tajima 等人在同一本杂志上报告，G-CSF 动员后的小鼠，多数干细胞位于 CD34⁺ 群体，这种表达同样可逆，并提出结果可能支持把动员血里的总 CD34⁺ 细胞作为植入预测指标、用 CD34 选择来富集人的干细胞。这些都是小鼠数据。在人体里「CD34⁻ 干细胞」和「CD34⁺ 干细胞」究竟是两种细胞，还是同一种细胞的两种状态，用免疫缺陷小鼠里的移植实验很难一锤定音。

走到病床

临床上的后果不抽象。2021 年，Frangoul 等人在 N Engl J Med 上报告了两例 BCL11A enhancer 的 CRISPR-Cas9 治疗：先对健康供者的 CD34⁺ 造血干/祖细胞做电穿孔，该位点约 80% 的等位基因被修饰，未见脱靶；随后，一名输血依赖型 β-地中海贫血患者和一名镰刀型贫血患者在清髓后接受了经同一方案编辑的自体 CD34⁺ 细胞。这里的 80% 是对整群 CD34⁺ 细胞（健康供者的）测得的等位基因修饰比例。它说明编辑这一步在群体层面做得相当充分，却没有告诉我们，在那极少数真正会终身产生血细胞的细胞里，编辑发生到了什么程度。平均数和少数细胞之间的距离，会在后面讲「效力」与「平均编辑率」的篇章里再展开。

传统看法认为，移植后的早期恢复由 committed 的 multipotent progenitor 驱动，随后才轮到长期植入的 HSC。2017 年，Radtke 等人在 Sci Transl Med 上报告，在非人灵长类的自体清髓移植模型里，他们观察到一群早期植入、却表现出 HSC 样行为并长期存在于体内的细胞；CD34⁺CD45RA⁻CD90⁺ 表型高度富集 HSC，这一群细胞能完整支持短期恢复和稳健的多谱系造血，并定量预测了移植的成败，以及中性粒细胞和血小板的恢复时间。所谓「CD34⁺ 剂量」，因此是一个粗的代理，细节在它底下那几个更细的子群里。

测量这些子群的实验，本身也带着时间维度。Notta 等人 2011 年的结果提醒我们，在免疫缺陷小鼠里，只能短暂植入的 MPP 与能产生长期多谱系移植物的 HSC 需要用 CD49f 才能分开。「谁是长期重建细胞」这个问题的答案，部分取决于用什么实验、观察多久去问。

还没解决的部分

人体内到底有多少造血是由 CD34⁻ 的干细胞贡献的，目前没有直接证据可以回答：现有的人源数据几乎都来自免疫缺陷小鼠。「真正的长期重建细胞」这个概念本身，也只能由移植实验在事后定义；表型再细，也仍是功能的代理人。而对基因治疗而言，要回答的问题始终是同一个：被改动的，是不是那一小撮会陪患者一辈子的细胞，还是整群混杂的 CD34⁺ 里那些更容易被编辑、却只管几周的细胞。

制服之所以流行，是因为它好认、好分拣、好计数，而临床确实需要这样的东西。只是当一件制服被当成身份证时，该去核实的就不再是标签，而是穿着它的人。

参考文献

  1. Civin CI, Strauss LC, Brovall C, Fackler MJ, Schwartz JF, Shaper JH. Antigenic analysis of hematopoiesis. III. A hematopoietic progenitor cell surface antigen defined by a monoclonal antibody raised against KG-1a cells. J Immunol. 1984;133(1):157-165. DOI
  2. Simmons DL, Satterthwaite AB, Tenen DG, Seed B. Molecular cloning of a cDNA encoding CD34, a sialomucin of human hematopoietic stem cells. J Immunol. 1992;148(1):267-271. DOI
  3. Berenson RJ, Andrews RG, Bensinger WI, et al. Antigen CD34+ marrow cells engraft lethally irradiated baboons. J Clin Invest. 1988;81(3):951-955. DOI
  4. Cheng J, Baumhueter S, Cacalano G, et al. Hematopoietic defects in mice lacking the sialomucin CD34. Blood. 1996;87(2):479-490. DOI
  5. Baum CM, Weissman IL, Tsukamoto AS, Buckle AM, Peault B. Isolation of a candidate human hematopoietic stem-cell population. Proc Natl Acad Sci U S A. 1992;89(7):2804-2808. DOI
  6. Majeti R, Park CY, Weissman IL. Identification of a hierarchy of multipotent hematopoietic progenitors in human cord blood. Cell Stem Cell. 2007;1(6):635-645. DOI
  7. 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):218-221. DOI
  8. Anjos-Afonso F, Buettner F, Mian SA, et al. Single cell analyses identify a highly regenerative and homogenous human CD34+ hematopoietic stem cell population. Nat Commun. 2022;13(1):2048. DOI
  9. Pellin D, Loperfido M, Baricordi C, et al. A comprehensive single cell transcriptional landscape of human hematopoietic progenitors. Nat Commun. 2019;10(1):2395. DOI
  10. Velten L, Haas SF, Raffel S, et al. Human haematopoietic stem cell lineage commitment is a continuous process. Nat Cell Biol. 2017;19(4):271-281. DOI
  11. Notta F, Zandi S, Takayama N, et al. Distinct routes of lineage development reshape the human blood hierarchy across ontogeny. Science. 2016;351(6269):aab2116. DOI
  12. Osawa M, Hanada K, Hamada H, Nakauchi H. Long-term lymphohematopoietic reconstitution by a single CD34-low/negative hematopoietic stem cell. Science. 1996;273(5272):242-245. DOI
  13. Bhatia M, Bonnet D, Murdoch B, Gan OI, Dick JE. A newly discovered class of human hematopoietic cells with SCID-repopulating activity. Nat Med. 1998;4(9):1038-1045. DOI
  14. Anjos-Afonso F, Currie E, Palmer HG, Foster KE, Taussig DC, Bonnet D. CD34(-) cells at the apex of the human hematopoietic stem cell hierarchy have distinctive cellular and molecular signatures. Cell Stem Cell. 2013;13(2):161-174. DOI
  15. Anjos-Afonso F, Bonnet D. Human CD34+ hematopoietic stem cell hierarchy: how far are we with its delineation at the most primitive level? Blood. 2023;142(6):509-518. DOI
  16. Sato T, Laver JH, Ogawa M. Reversible expression of CD34 by murine hematopoietic stem cells. Blood. 1999;94(8):2548-2554. DOI
  17. Tajima F, Sato T, Laver JH, Ogawa M. CD34 expression by murine hematopoietic stem cells mobilized by granulocyte colony-stimulating factor. Blood. 2000;96(5):1989-1993. DOI
  18. Frangoul H, Altshuler D, Cappellini MD, et al. CRISPR-Cas9 Gene Editing for Sickle Cell Disease and β-Thalassemia. N Engl J Med. 2021;384(3):252-260. DOI
  19. Radtke S, Adair JE, Giese MA, et al. A distinct hematopoietic stem cell population for rapid multilineage engraftment in nonhuman primates. Sci Transl Med. 2017;9(414):eaan1145. DOI

In descriptions of the starting material for autologous hematopoietic stem cell transplantation and gene-editing therapies, “CD34⁺ cells” is a phrase that is hard to avoid. Over time, the count has come to be read as how many stem cells there are. But CD34 is not an ID card engraved with “I am a hematopoietic stem cell”. It is closer to a uniform widely worn in the marrow: among those wearing it are the very few who sustain blood production for a lifetime, a large crowd of intermediate players who only handle the next few weeks, and, occasionally, someone who has changed into street clothes, uniform absent, yet still able to carry the entire blood system.

A handy handle

The story begins with an antibody. In 1984, Civin and colleagues reported in J Immunol the antibody My-10: it was raised against the immature myeloid cell line KG-1a and selected for not reacting with mature granulocytes, and it was found to bind specifically to immature normal human marrow cells, including hematopoietic progenitor cells; the authors therefore pointed out that the antibody could be used to identify and purify hematopoietic progenitors. In 1992, Simmons and colleagues, in the same journal, cloned the CD34 cDNA from a KG-1 cell library and described it as a transmembrane glycoprotein of about 115 kDa, with unknown function, found not only on hematopoietic progenitors but also on the endothelium of small blood vessels in many tissues.

What turned this antigen from a marker into a tool was a transplant experiment. In 1988, Berenson and colleagues reported in J Clin Invest that CD34 is present on 1–4% of human marrow cells and covers nearly all hematopoietic progenitors detectable in vitro. Using the reactive antibody 12-8, they enriched autologous marrow from five baboons; of the 15–27 × 10⁶ cells after enrichment, 65–91% expressed CD34. After lethal irradiation, the animals were infused with these cells, and all five had granulocyte counts above 1,000/mm³ and platelets above 20 × 10³/mm³ within 13–24 days, after which blood counts returned to normal. As controls, two other animals received marrow depleted of CD34⁺ cells, at 184 and 285 × 10⁶ cells per kilogram of body weight (versus 3.2–4.4 × 10⁶ in the former group); one died on day 29 without engrafting, and the other remained pancytopenic for more than 100 days. The authors’ conclusion was restrained: the data suggest that the stem cells responsible for hematopoietic reconstitution are CD34⁺.

For a clinical field that wanted to separate stem cells from marrow and use them, this was enough. An antibody could grab them, a count could quantify them, and a transplant could validate them, so CD34 became a standard step in the workflow. What also needs to be remembered is that it is a handle, not a component of stemness itself: in 1996, Cheng and colleagues reported in Blood that mice lacking CD34 developed normally, had adult blood counts that looked no different from wild type, and after sublethal irradiation showed erythroid, myeloid and platelet recovery kinetics indistinguishable from wild type, even though colony-forming activity of progenitors in marrow and spleen was clearly reduced; the authors also noted that CD34 plays an important role in the formation of progenitors, and that adult mice may simply retain a sufficient reserve of progenitors. At least for adult steady state and recovery after sublethal irradiation, mice without it showed no difference in blood counts.

Beneath the badge is a layered population

A handy handle is not the same as a homogeneous catch. From the 1990s onward, almost every purification of human hematopoietic stem cells has amounted to subtraction inside the CD34⁺ compartment. In 1992, Baum and colleagues in PNAS gated cells expressing Thy-1 and CD34 and lacking known lineage markers from human fetal bone marrow; they made up only 0.05–0.1%, yet concentrated the activity that could establish long-term multilineage cultures on mouse stromal cells. In 2007, Majeti and colleagues in Cell Stem Cell split the Lin⁻CD34⁺CD38⁻ fraction of cord blood and marrow into three groups by CD90 and CD45RA: in cord blood the CD90⁺CD45RA⁻ group contained HSCs, with transplantation activity that could be narrowed down to as few as 10 purified cells, while the CD90⁻CD45RA⁻ group was prospectively isolated for the first time as a candidate multipotent progenitor (MPP). In 2011, Notta and colleagues in Science went further, using CD49f to distinguish HSCs from MPPs capable of only transient engraftment: a single CD49f⁺ cell could efficiently generate a long-term multilineage graft, whereas cells that lost CD49f engrafted only transiently. By 2022, Anjos-Afonso and colleagues in Nat Commun reported a CD34⁺EPCR⁺(CD38/CD45RA)⁻ population whose stem cell frequency in xenotransplantation was about 1 in 3 cells, which the authors described as the highest value described to date.

From 1–4% of marrow cells all the way down to one in three says the same thing: CD34⁺ is an entrance, not an endpoint. And the structure inside that entrance may not be a neat row of steps. In 2019, Pellin and colleagues in Nat Commun used single-cell RNA-seq to test existing sorting schemes and found that CD34⁺ marrow cells cut by these schemes do correspond to a hierarchical transcriptional landscape, but that this landscape misses important early fate decisions; they also proposed CD164 as a candidate marker of the earliest branching. In 2017, Velten and colleagues in Nat Cell Biol further reported that, at steady state, individual HSCs gradually acquire lineage biases, and lineage-restricted cells emerge directly from a low-primed undifferentiated continuum rather than passing through discrete progenitor steps in sequence. In 2016, Notta and colleagues in Science, starting from single CD34⁺ cells, mapped the fate of hematopoiesis at different developmental stages: fetal liver contains abundant oligopotent progenitors with mixed myeloid, erythroid and megakaryocyte fates, whereas adult marrow is left mainly with multipotent and unipotent types. Under the same marker, what it contains changes with developmental stage. Whether the blood tree is a discrete hierarchy or a continuous landscape is a debate for another installment; for now, it is enough to remember one thing: CD34⁺ is a mixed population that has been cut over and over by various sorting schemes.

Cells without the badge

More troublesome are the examples in the opposite direction. In 1996, Osawa and colleagues found in Science that, unlike in humans, primitive HSC activity in adult mouse marrow resides in the CD34 low-to-negative fraction; injecting a single CD34-low or -negative, c-Kit⁺, Sca-1⁺, lineage-marker-negative cell achieved long-term reconstitution of the lymphohematopoietic system in 21% of recipients. In 1998, Bhatia and colleagues reported a similar phenomenon in human cells in Nat Med: in NOD/SCID mice there is a class of human cells able to initiate multilineage hematopoiesis, which they called CD34neg-SRC; these are confined to the Lin⁻CD34⁻CD38⁻ population, have very low colony-forming ability in vitro, and increase in number under certain short-term suspension culture conditions, while the same conditions could not maintain SRCs derived from the CD34⁺ population. In 2013, Anjos-Afonso and colleagues in Cell Stem Cell compared the Lin⁻CD34⁻CD38⁻CD93hi population, argued that it contains HSCs, and placed them above the CD34⁺ population, in a rather primitive and quiescent state maintained by signals such as Notch.

These cells are all rare. In 2023, in a Blood review, Anjos-Afonso and Bonnet from the same laboratory mentioned that an extremely rare CD34⁻ HSC population is thought to sit above CD34⁺ HSCs, while also writing that human hematopoiesis is mainly sustained by the latter. So the more accurate statement is not that CD34⁻ cells are the true stem cells, but that the CD34 badge both lets in cells that are not stem cells and, occasionally, keeps out cells that are.

There is another reading of how CD34 varies in mice, one closer to state than to identity. In 1999, Sato and colleagues in Blood used a mouse transplantation model to report that in normal adult mice, most stem cells with long-term engraftment capacity are CD34⁻; after 5-FU treatment, stem cells can be found in both the CD34⁻ and CD34⁺ fractions; when lineage-negative, c-Kit⁺, Sca-1⁺, CD34⁻ marrow cells were expanded in vitro for one week in IL-11 and steel factor, only the CD34⁺ fraction could still engraft over 2 to 5 months; and when CD34⁺ cells from after 5-FU were transplanted, once the marrow returned to steady state, most cells with long-term engraftment capacity were again in the CD34⁻ fraction. The authors therefore concluded that CD34 expression reflects the activation state of stem cells, and that it is reversible. In 2000, Tajima and colleagues in the same journal reported that in G-CSF-mobilized mice most stem cells reside in the CD34⁺ population, that this expression is likewise reversible, and proposed that the results may support using total CD34⁺ cells in mobilized blood as a predictor of engraftment, and using CD34 selection to enrich human stem cells. These are all mouse data. In humans, whether CD34⁻ stem cells and CD34⁺ stem cells are two kinds of cells or two states of one kind is hard to settle conclusively with transplantation experiments in immunodeficient mice.

Reaching the bedside

The clinical consequences are not abstract. In 2021, Frangoul and colleagues reported in N Engl J Med two cases of CRISPR-Cas9 editing of the BCL11A enhancer: CD34⁺ hematopoietic stem and progenitor cells from healthy donors were first electroporated, about 80% of alleles at the site were modified, and no off-target editing was seen; then a patient with transfusion-dependent β-thalassemia and a patient with sickle cell disease received, after myeloablation, autologous CD34⁺ cells edited with the same protocol. The 80% here is the allele modification rate measured across the whole CD34⁺ population (from healthy donors). It shows that the editing step is done quite thoroughly at the population level, but it does not tell us how far editing reached within the very few cells that will actually produce blood for life. The distance between the average and the few will be taken up in later installments on potency and average editing rates.

The traditional view holds that early recovery after transplantation is driven by committed multipotent progenitors, with long-term engrafting HSCs taking over afterwards. In 2017, Radtke and colleagues reported in Sci Transl Med that in a nonhuman primate model of autologous myeloablative transplantation, they observed a group of cells that engrafted early yet behaved like HSCs and persisted long term in the body; the CD34⁺CD45RA⁻CD90⁺ phenotype was highly enriched for HSCs, and this group of cells fully supported short-term recovery and robust multilineage hematopoiesis, and quantitatively predicted transplant success and the time to neutrophil and platelet recovery. The so-called CD34⁺ dose is therefore a coarse proxy; the details lie in the finer subgroups beneath it.

The experiments that measure these subgroups also carry a time dimension. The 2011 result of Notta and colleagues reminds us that, in immunodeficient mice, MPPs capable of only transient engraftment and HSCs capable of producing long-term multilineage grafts need CD49f to be told apart. The answer to who is a long-term reconstituting cell depends in part on which experiment is used and how long it is observed.

What remains unresolved

How much of human hematopoiesis is contributed by CD34⁻ stem cells has no direct evidence to answer it at present: nearly all existing human data come from immunodeficient mice. The very concept of a true long-term reconstituting cell can also only be defined after the fact by transplantation; however fine the phenotype, it remains a proxy for function. And for gene therapy, the question to answer is always the same: is what was altered the small handful of cells that will stay with the patient for life, or the cells within the mixed CD34⁺ population that are easier to edit but only serve for a few weeks?

The uniform became popular because it is easy to recognize, easy to sort and easy to count, and the clinic does need such a thing. But once a uniform is treated as an ID card, what needs checking is no longer the label but the person wearing it.

References

  1. Civin CI, Strauss LC, Brovall C, Fackler MJ, Schwartz JF, Shaper JH. Antigenic analysis of hematopoiesis. III. A hematopoietic progenitor cell surface antigen defined by a monoclonal antibody raised against KG-1a cells. J Immunol. 1984;133(1):157-165. DOI
  2. Simmons DL, Satterthwaite AB, Tenen DG, Seed B. Molecular cloning of a cDNA encoding CD34, a sialomucin of human hematopoietic stem cells. J Immunol. 1992;148(1):267-271. DOI
  3. Berenson RJ, Andrews RG, Bensinger WI, et al. Antigen CD34+ marrow cells engraft lethally irradiated baboons. J Clin Invest. 1988;81(3):951-955. DOI
  4. Cheng J, Baumhueter S, Cacalano G, et al. Hematopoietic defects in mice lacking the sialomucin CD34. Blood. 1996;87(2):479-490. DOI
  5. Baum CM, Weissman IL, Tsukamoto AS, Buckle AM, Peault B. Isolation of a candidate human hematopoietic stem-cell population. Proc Natl Acad Sci U S A. 1992;89(7):2804-2808. DOI
  6. Majeti R, Park CY, Weissman IL. Identification of a hierarchy of multipotent hematopoietic progenitors in human cord blood. Cell Stem Cell. 2007;1(6):635-645. DOI
  7. 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):218-221. DOI
  8. Anjos-Afonso F, Buettner F, Mian SA, et al. Single cell analyses identify a highly regenerative and homogenous human CD34+ hematopoietic stem cell population. Nat Commun. 2022;13(1):2048. DOI
  9. Pellin D, Loperfido M, Baricordi C, et al. A comprehensive single cell transcriptional landscape of human hematopoietic progenitors. Nat Commun. 2019;10(1):2395. DOI
  10. Velten L, Haas SF, Raffel S, et al. Human haematopoietic stem cell lineage commitment is a continuous process. Nat Cell Biol. 2017;19(4):271-281. DOI
  11. Notta F, Zandi S, Takayama N, et al. Distinct routes of lineage development reshape the human blood hierarchy across ontogeny. Science. 2016;351(6269):aab2116. DOI
  12. Osawa M, Hanada K, Hamada H, Nakauchi H. Long-term lymphohematopoietic reconstitution by a single CD34-low/negative hematopoietic stem cell. Science. 1996;273(5272):242-245. DOI
  13. Bhatia M, Bonnet D, Murdoch B, Gan OI, Dick JE. A newly discovered class of human hematopoietic cells with SCID-repopulating activity. Nat Med. 1998;4(9):1038-1045. DOI
  14. Anjos-Afonso F, Currie E, Palmer HG, Foster KE, Taussig DC, Bonnet D. CD34(-) cells at the apex of the human hematopoietic stem cell hierarchy have distinctive cellular and molecular signatures. Cell Stem Cell. 2013;13(2):161-174. DOI
  15. Anjos-Afonso F, Bonnet D. Human CD34+ hematopoietic stem cell hierarchy: how far are we with its delineation at the most primitive level? Blood. 2023;142(6):509-518. DOI
  16. Sato T, Laver JH, Ogawa M. Reversible expression of CD34 by murine hematopoietic stem cells. Blood. 1999;94(8):2548-2554. DOI
  17. Tajima F, Sato T, Laver JH, Ogawa M. CD34 expression by murine hematopoietic stem cells mobilized by granulocyte colony-stimulating factor. Blood. 2000;96(5):1989-1993. DOI
  18. Frangoul H, Altshuler D, Cappellini MD, et al. CRISPR-Cas9 Gene Editing for Sickle Cell Disease and β-Thalassemia. N Engl J Med. 2021;384(3):252-260. DOI
  19. Radtke S, Adair JE, Giese MA, et al. A distinct hematopoietic stem cell population for rapid multilineage engraftment in nonhuman primates. Sci Transl Med. 2017;9(414):eaan1145. DOI