Yang Liu

← Primer

HSC · Healing Source Code

预处理不是一件事——清病、压免疫、腾龛位该分开看 Conditioning is not one job — clearing disease, suppressing immunity and making room should be read separately

2026-10-09 · conditioning, busulfan, niche, reduced-intensity conditioning, HSC transplantation

同一个词 conditioning,会出现在三种很不一样的病历上:一位要清掉白血病的成人;一位需要被压住免疫、才不会排斥供者细胞的受者;还有一位镰刀型贫血患者,他只想让自己那批编辑过的造血干细胞在骨髓里住下来。三份病历写着同一个词,要它完成的任务却并不相同。把这个词拆开,是理解移植的起点。

两个老目的,和一把量骨髓损伤的尺子

Bacigalupo 等人 2009 年在 Biol Blood Marrow Transplant 上提出 conditioning 强度的工作定义。文章开篇说,异体移植前用化疗或化疗加放疗预处理,有两个目的:疾病是肿瘤时减少肿瘤负荷,以及压制受者的免疫系统,让干细胞得以植入。文中点出的例外是重症联合免疫缺陷(SCID)的婴儿,以及有同卵双胞胎供者的重型再生障碍性贫血患者,可以不做 conditioning 就接受移植。

早期方案靠剂量强度。Gyurkocza 与 Sandmaier 2014 年在 Blood 上回顾说,后来人们认识到 graft-versus-tumor 效应的贡献,许多研究者开始降低剂量,这被他们称为一次重大的范式转变,可接受移植的患者群体也随之明显扩大。

Bacigalupo 的定义按血细胞减少持续多久、是否必须有干细胞支持分类:清髓(MA)造成不可逆的血细胞减少,必须有干细胞支持;非清髓(NMA)只造成轻微的血细胞减少;减强度(RIC)介于两者之间。这把尺子量的是骨髓被伤到什么程度,并没有回答 conditioning 究竟在替你做哪件事。

清病:强度确实在起作用

2017 年,Scott 等人在 J Clin Oncol 上发表了 BMT CTN 0901 的结果(NCT01339910):272 名 18–65 岁、移植前骨髓原始细胞低于 5% 的 AML 或 MDS 患者,被随机分到清髓或减强度方案,供者均为 HLA 匹配。试验原计划入组 356 人,因减强度组复发率过高而在 272 人时停止——减强度组 48.3%,清髓组 13.5%。反过来,移植相关死亡率是减强度组 4.4%,清髓组 15.8%。18 个月总生存率为 67.7% 对 77.5%,差异没有达到统计学显著(P = .07),无复发生存率则是清髓组显著更高(67.8% 对 47.3%)。在这批患者里,强度换来的是对疾病更好的控制,代价是更多的毒性死亡。

压免疫:不一定要把骨髓清空

第二件事,是不让受者的免疫系统把供者细胞赶走。1998 年,Slavin 等人在 Blood 上报告了 26 名患者(多数为恶性血液病,4 人为遗传病):用 fludarabine、抗 T 淋巴细胞球蛋白和低剂量 busulfan(8 mg/kg)预处理,出发点是建立 host-versus-graft 耐受。结果 9 人形成稳定的部分嵌合、17 人为完全嵌合。

镰刀型贫血的成人数据更直观。2014 年,Hsieh 等人在 JAMA 上报告了 30 名 16–65 岁的重症患者(NCT00061568),方案是 alemtuzumab、300 cGy 全身照射和 sirolimus,移植物来自 HLA 匹配的同胞。26 名患者(87%)得到长期稳定的供者植入,没有急性或慢性 GVHD;平均供者髓系嵌合率 86%,供者 T 细胞 48%。这是一项单臂研究,方案属于低强度,多数人仍然植入了。

腾龛位:植入的另一道门

Bacigalupo 的第二个目的”让干细胞得以植入”里,其实可以分出两道门:受者的免疫,以及骨髓里的位置。后一道门的证据来自 2007 年 Czechowicz 等人在 Science 上的小鼠实验:除免疫屏障之外,供者 HSC 的植入还受限于宿主 HSC 对合适龛位的占据。在免疫缺陷小鼠里,阻断 c-kit 功能的抗体 ACK2 可短暂清除超过 98% 的内源 HSC;随后输入供者 HSC,嵌合率最高可达 90%。这是临床前证据;作者在摘要末尾只说,把方法外推到人类,或许能带来温和而有效的 conditioning。

回到非恶性病的基因治疗:没有肿瘤要清,这一点是明确的。至于免疫这一环,本篇引用的文献没有就自体细胞回输是否需要压制免疫给出专门论述,这里不下判断。能从文献里直接读到的是:Sharma 与 Corbacioglu 2026 年在 Blood Adv 上的综述,把 conditioning 在血红蛋白病基因治疗里的核心作用概括为使基因修饰的造血干细胞得以持久植入,并指出不足的龛位清除与过度毒性之间空间很窄。试验里常用的 conditioning 是 busulfan,例如 Thompson 等人 2018 年在 N Engl J Med 上报告的 β-地中海贫血 LentiGlobin 试验(清髓 busulfan 预处理);Kwiatkowski 等人 2024 年在 Lancet 上报告的 HGB-212 采用 busulfan 基础上、经药代动力学调整的清髓方案(NCT03207009)。

Aiuti 等人 2009 年在 N Engl J Med 上治疗 10 名 ADA 缺乏型 SCID 儿童,只用了含 busulfan 的非清髓预处理(NCT00598481、NCT00599781):1 年时骨髓髓系细胞中含 ADA 的比例平均约 3.5%–8.9%,外周血淋巴细胞为 52.4%–88.0%,9 名患者实现免疫重建。不同疾病、不同方案究竟需要腾出多少位置,这些研究没有直接比较。

busulfan 的账,以及监测剂量

窗口的另一端是毒性。Cattoni 等人 2023 年在 Front Endocrinol 上对一家中心 1992–2021 年间儿童期接受 HSCT 的 130 名男性做了回顾性分析:发生不同程度性腺功能障碍的比例,在以全身照射为主的方案后是 85%,busulfan 方案后是 51%,环磷酰胺方案后是 32%,treosulfan 方案后为 0%。这是单中心回顾性数据,血红蛋白病仅 6 例,并非基因治疗队列,不宜直接套到基因治疗患者身上。基因治疗侧目前能引的多是个案:Yin 等人 2026 年在 Blood Adv 上报告的一位 14 岁 α-地中海贫血女孩,接受 busulfan 预处理后的不良事件限于预期的 busulfan 相关毒性,包括 2 级月经不规律和被列为 serious AE 的卵巢功能衰竭。Sharma 与 Corbacioglu 的综述把不育与潜在的基因毒性,列为 busulfan 的急性之外的长期代价。

既然要用,就要监测暴露量。Bartelink 等人 2016 年在 Lancet Haematol 上回顾了来自 15 个中心的 674 名儿童与青年异体移植患者(41% 为恶性病,59% 为非恶性病):静脉 busulfan 累积 AUC 落在 78–101 mg × h/L 的患者,2 年无事件生存率为 77.0%;落在 58–86 mg × h/L 的历史低目标区间为 66.1%;高于 101 的只有 49.5%(P = .011)。AUC 过高伴随更多的急性毒性和移植相关死亡。作者认为用经过验证的药代动力学模型把 AUC 瞄准 78–101 有望改善结局。这是回顾性的异体移植人群分析,不是基因治疗数据。

基因治疗这边也在做类似的监测。Canarutto 等人 2026 年在 Front Pharmacol 上回顾了 54 名接受每日一次静脉 busulfan 并做治疗药物监测的儿童:累积 AUC 虽都控制在目标范围内,单次剂量的峰值 AUC 越高,黏膜炎越长、肠外营养需求越多;起始剂量 80 mg/m² 比 120 mg/m² 的峰值 AUC 更低、后续调量更少,且植入结局没有差异。

还没解决的部分

三件事拆开之后,空白也就清楚了。腾龛位到底要腾出多少,目前主要只能从疗效反推;上文的数字多来自异体移植或小样本,基因治疗里 busulfan 暴露与长期植入、不育之间的关系仍缺系统数据。Sharma 与 Corbacioglu 也认为,优化需要对比试验和长期随访。

回到开头那三份病历。同一个词之下,有一件是为了清病,有一件是为了压免疫,有一件是为了让位置空出来。Scott 的随机试验、Hsieh 的低强度移植、Czechowicz 的小鼠实验,分别碰到了其中一件,证据的层级也各不相同。基因治疗要做的主要是哪一件,能不能不用同一把推土机来做,是拆开之后才问得出的问题。

参考文献

  1. Bacigalupo A, Ballen K, Rizzo D, et al. Defining the intensity of conditioning regimens: working definitions. Biol Blood Marrow Transplant. 2009;15(12):1628-1633. DOI PMID: 19896087
  2. Gyurkocza B, Sandmaier BM. Conditioning regimens for hematopoietic cell transplantation: one size does not fit all. Blood. 2014;124(3):344-353. DOI PMID: 24914142
  3. Scott BL, Pasquini MC, Logan BR, et al. Myeloablative versus reduced-intensity hematopoietic cell transplantation for acute myeloid leukemia and myelodysplastic syndromes. J Clin Oncol. 2017;35(11):1154-1161. DOI PMID: 28380315(NCT01339910)
  4. Slavin S, Nagler A, Naparstek E, et al. Nonmyeloablative stem cell transplantation and cell therapy as an alternative to conventional bone marrow transplantation with lethal cytoreduction for the treatment of malignant and nonmalignant hematologic diseases. Blood. 1998;91(3):756-763. DOI PMID: 9446633
  5. Hsieh MM, Fitzhugh CD, Weitzel RP, et al. Nonmyeloablative HLA-matched sibling allogeneic hematopoietic stem cell transplantation for severe sickle cell phenotype. JAMA. 2014;312(1):48-56. DOI PMID: 25058217(NCT00061568)
  6. Czechowicz A, Kraft D, Weissman IL, Bhattacharya D. Efficient transplantation via antibody-based clearance of hematopoietic stem cell niches. Science. 2007;318(5854):1296-1299. DOI PMID: 18033883
  7. Thompson AA, Walters MC, Kwiatkowski J, et al. Gene therapy in patients with transfusion-dependent β-thalassemia. N Engl J Med. 2018;378(16):1479-1493. DOI PMID: 29669226
  8. Kwiatkowski JL, Walters MC, Hongeng S, et al. Betibeglogene autotemcel gene therapy in patients with transfusion-dependent, severe genotype β-thalassaemia (HGB-212): a non-randomised, multicentre, single-arm, open-label, single-dose, phase 3 trial. Lancet. 2024;404(10468):2175-2186. DOI PMID: 39527960(NCT03207009)
  9. Aiuti A, Cattaneo F, Galimberti S, et al. Gene therapy for immunodeficiency due to adenosine deaminase deficiency. N Engl J Med. 2009;360(5):447-458. DOI PMID: 19179314(NCT00598481, NCT00599781)
  10. Sharma A, Corbacioglu S. Alternative conditioning regimens for hemoglobinopathy gene therapy: balancing efficacy, toxicity, and the next frontier. Blood Adv. 2026;10(18):6118-6123. DOI PMID: 42418372
  11. Cattoni A, Nicolosi ML, Capitoli G, et al. Pubertal attainment and Leydig cell function following pediatric hematopoietic stem cell transplantation: a three-decade longitudinal assessment. Front Endocrinol. 2023;14:1292683. DOI PMID: 38152128
  12. Yin X, Qian L, Dong G, et al. Successful gene therapy for transfusion-dependent α-thalassemia: a case report. Blood Adv. 2026;10(14):5153-5160. DOI PMID: 42200456
  13. Bartelink IH, Lalmohamed A, van Reij EM, et al. Association of busulfan exposure with survival and toxicity after haemopoietic cell transplantation in children and young adults: a multicentre, retrospective cohort analysis. Lancet Haematol. 2016;3(11):e526-e536. DOI PMID: 27746112
  14. Canarutto D, De Gregori S, Recupero S, et al. Busulfan exposure variability and acute toxicity in pediatric patients undergoing hematopoietic stem cell gene therapy. Front Pharmacol. 2026;17:1872108. DOI PMID: 42568941

The same word, conditioning, turns up on three very different charts: an adult who needs leukemia cleared; a recipient whose immune system has to be held down so it will not reject donor cells; and a patient with sickle cell disease who only wants their own edited hematopoietic stem cells to take up residence in the marrow. Three charts, one word, and the job it is asked to do is not the same. Taking the word apart is where understanding transplantation begins.

Two old purposes, and a ruler for marrow damage

In 2009, Bacigalupo and colleagues proposed working definitions of conditioning intensity in Biol Blood Marrow Transplant. The article opens by saying that conditioning with chemotherapy, or chemotherapy plus radiation, before allogeneic transplantation has two purposes: reducing tumor burden when the disease is a malignancy, and suppressing the recipient’s immune system so that the stem cells can engraft. The exceptions it points out are infants with severe combined immunodeficiency (SCID) and patients with severe aplastic anemia who have an identical-twin donor, who can receive a transplant without conditioning.

Early regimens relied on dose intensity. Reviewing the field in Blood in 2014, Gyurkocza and Sandmaier wrote that once the contribution of the graft-versus-tumor effect was appreciated, many investigators began lowering doses; they called this a major paradigm shift, and the population of patients who could be transplanted expanded markedly as a result.

Bacigalupo’s definition classifies regimens by how long cytopenia lasts and whether stem cell support is mandatory: myeloablative (MA) conditioning causes irreversible cytopenia and requires stem cell support; nonmyeloablative (NMA) conditioning causes minimal cytopenia; reduced-intensity conditioning (RIC) falls in between. This ruler measures how badly the marrow is hurt. It does not say which job the conditioning is actually doing for you.

Clearing disease: intensity does matter

In 2017, Scott and colleagues published the results of BMT CTN 0901 in J Clin Oncol (NCT01339910): 272 patients aged 18–65 with AML or MDS and fewer than 5% marrow blasts before transplant were randomized to a myeloablative or a reduced-intensity regimen, all with HLA-matched donors. The trial had planned to enroll 356 and was stopped at 272 because relapse was too high in the reduced-intensity arm: 48.3% versus 13.5% in the myeloablative arm. In the other direction, transplant-related mortality was 4.4% in the reduced-intensity arm and 15.8% in the myeloablative arm. Overall survival at 18 months was 67.7% versus 77.5%, a difference that did not reach statistical significance (P = .07), while relapse-free survival was significantly higher with myeloablative conditioning (67.8% versus 47.3%). In these patients, intensity bought better control of the disease, at the price of more deaths from toxicity.

Suppressing immunity: the marrow need not be emptied

The second job is to keep the recipient’s immune system from driving out the donor cells. In 1998, Slavin and colleagues reported in Blood on 26 patients (most with hematologic malignancies, four with genetic diseases) conditioned with fludarabine, antithymocyte globulin and low-dose busulfan (8 mg/kg), an approach built on establishing host-versus-graft tolerance. Nine patients developed stable partial chimerism and 17 full chimerism.

The data in adults with sickle cell disease are more direct. In 2014, Hsieh and colleagues reported in JAMA on 30 patients aged 16–65 with severe disease (NCT00061568), conditioned with alemtuzumab, 300 cGy total body irradiation and sirolimus, with grafts from HLA-matched siblings. Twenty-six patients (87%) achieved long-term stable donor engraftment, with no acute or chronic GVHD; mean donor myeloid chimerism was 86% and donor T-cell chimerism 48%. This was a single-arm study with a low-intensity regimen, and most people still engrafted.

Making room: another door to engraftment

Bacigalupo’s second purpose, letting the stem cells engraft, can in fact be split into two doors: the recipient’s immunity, and the physical space in the marrow. The evidence for the latter comes from a 2007 mouse study by Czechowicz and colleagues in Science: beyond the immune barrier, engraftment of donor HSCs is also limited by occupancy of suitable niches by host HSCs. In immunodeficient mice, ACK2, an antibody that blocks c-kit function, transiently depleted more than 98% of endogenous HSCs; donor HSCs infused afterward reached chimerism of up to 90%. This is preclinical evidence. At the end of the abstract the authors say only that extending the approach to humans might allow gentle yet effective conditioning.

Back to gene therapy for non-malignant disease: there is no tumor to clear, and that much is clear. As for the immune side, the literature cited in this piece offers no dedicated discussion of whether autologous cell infusions need immune suppression, so no judgment is made here. What can be read directly from the literature is this: in a 2026 review in Blood Adv, Sharma and Corbacioglu summarize the central role of conditioning in hemoglobinopathy gene therapy as enabling durable engraftment of gene-modified HSCs, and note the narrow space between inadequate niche clearance and excessive toxicity. The conditioning commonly used in trials is busulfan: Thompson and colleagues reported the LentiGlobin trial in β-thalassemia in N Engl J Med in 2018 (myeloablative busulfan conditioning), and Kwiatkowski and colleagues reported HGB-212 in Lancet in 2024, which used a pharmacokinetically adjusted myeloablative regimen built on busulfan (NCT03207009).

In 2009, Aiuti and colleagues treated 10 children with ADA-deficient SCID in N Engl J Med using only nonmyeloablative conditioning containing busulfan (NCT00598481, NCT00599781): at one year, the proportion of ADA-containing cells in bone marrow myeloid cells averaged roughly 3.5%–8.9%, and in peripheral blood lymphocytes 52.4%–88.0%, and nine patients achieved immune reconstitution. How much room different diseases and different regimens actually need was not directly compared in these studies.

The cost of busulfan, and monitoring the dose

The other end of the window is toxicity. In 2023, Cattoni and colleagues in Front Endocrinol retrospectively analyzed 130 males who underwent HSCT in childhood at a single center between 1992 and 2021: the proportion with some degree of gonadal dysfunction was 85% after total-body-irradiation-based regimens, 51% after busulfan regimens, 32% after cyclophosphamide regimens, and 0% after treosulfan regimens. These are single-center retrospective data, only six patients had hemoglobinopathies, and this was not a gene therapy cohort, so they should not be applied directly to gene therapy patients. On the gene therapy side, what can currently be cited is mostly case reports: in a 14-year-old girl with α-thalassemia reported by Yin and colleagues in Blood Adv in 2026, adverse events after busulfan conditioning were limited to expected busulfan-related toxicities, including grade 2 irregular menses and ovarian failure classified as a serious AE. The Sharma and Corbacioglu review lists infertility and potential genotoxicity among busulfan’s costs beyond the acute ones.

If it is going to be used, exposure has to be monitored. In 2016, Bartelink and colleagues reviewed 674 children and young adults from 15 centers who underwent allogeneic transplantation (41% with malignant disease, 59% with non-malignant disease) in Lancet Haematol: among patients whose cumulative intravenous busulfan AUC fell within 78–101 mg × h/L, 2-year event-free survival was 77.0%; within the historical lower target range of 58–86 mg × h/L it was 66.1%; above 101 it was only 49.5% (P = .011). Excessive AUC went along with more acute toxicity and transplant-related mortality. The authors suggested that using a validated pharmacokinetic model to target an AUC of 78–101 may improve outcomes. This was a retrospective analysis in an allogeneic transplant population, not gene therapy data.

Similar monitoring is being done on the gene therapy side. In Front Pharmacol in 2026, Canarutto and colleagues reviewed 54 children who received once-daily intravenous busulfan with therapeutic drug monitoring: although cumulative AUC was kept within the target range in all of them, the higher the peak single-dose AUC, the longer the mucositis and the greater the need for parenteral nutrition; a starting dose of 80 mg/m² gave a lower peak AUC than 120 mg/m², with fewer subsequent dose adjustments and no difference in engraftment outcomes.

What remains unresolved

With the three jobs pulled apart, the gaps become clear. How much niche space actually needs to be cleared can for now mostly be inferred backward from efficacy; the numbers above come largely from allogeneic transplantation or small samples, and systematic data linking busulfan exposure in gene therapy to long-term engraftment and infertility are still lacking. Sharma and Corbacioglu likewise consider that optimization will require comparative trials and long-term follow-up.

Back to the three charts at the start. Under one word, one job is to clear disease, one is to suppress immunity, and one is to leave the space open. Scott’s randomized trial, Hsieh’s low-intensity transplant and Czechowicz’s mouse experiment each touch one of them, and the levels of evidence differ. Which of the three gene therapy mainly needs, and whether it can be done without using the same bulldozer for all of them, are questions that can only be asked once the word has been taken apart.

References

  1. Bacigalupo A, Ballen K, Rizzo D, et al. Defining the intensity of conditioning regimens: working definitions. Biol Blood Marrow Transplant. 2009;15(12):1628-1633. DOI PMID: 19896087
  2. Gyurkocza B, Sandmaier BM. Conditioning regimens for hematopoietic cell transplantation: one size does not fit all. Blood. 2014;124(3):344-353. DOI PMID: 24914142
  3. Scott BL, Pasquini MC, Logan BR, et al. Myeloablative versus reduced-intensity hematopoietic cell transplantation for acute myeloid leukemia and myelodysplastic syndromes. J Clin Oncol. 2017;35(11):1154-1161. DOI PMID: 28380315(NCT01339910)
  4. Slavin S, Nagler A, Naparstek E, et al. Nonmyeloablative stem cell transplantation and cell therapy as an alternative to conventional bone marrow transplantation with lethal cytoreduction for the treatment of malignant and nonmalignant hematologic diseases. Blood. 1998;91(3):756-763. DOI PMID: 9446633
  5. Hsieh MM, Fitzhugh CD, Weitzel RP, et al. Nonmyeloablative HLA-matched sibling allogeneic hematopoietic stem cell transplantation for severe sickle cell phenotype. JAMA. 2014;312(1):48-56. DOI PMID: 25058217(NCT00061568)
  6. Czechowicz A, Kraft D, Weissman IL, Bhattacharya D. Efficient transplantation via antibody-based clearance of hematopoietic stem cell niches. Science. 2007;318(5854):1296-1299. DOI PMID: 18033883
  7. Thompson AA, Walters MC, Kwiatkowski J, et al. Gene therapy in patients with transfusion-dependent β-thalassemia. N Engl J Med. 2018;378(16):1479-1493. DOI PMID: 29669226
  8. Kwiatkowski JL, Walters MC, Hongeng S, et al. Betibeglogene autotemcel gene therapy in patients with transfusion-dependent, severe genotype β-thalassaemia (HGB-212): a non-randomised, multicentre, single-arm, open-label, single-dose, phase 3 trial. Lancet. 2024;404(10468):2175-2186. DOI PMID: 39527960(NCT03207009)
  9. Aiuti A, Cattaneo F, Galimberti S, et al. Gene therapy for immunodeficiency due to adenosine deaminase deficiency. N Engl J Med. 2009;360(5):447-458. DOI PMID: 19179314(NCT00598481, NCT00599781)
  10. Sharma A, Corbacioglu S. Alternative conditioning regimens for hemoglobinopathy gene therapy: balancing efficacy, toxicity, and the next frontier. Blood Adv. 2026;10(18):6118-6123. DOI PMID: 42418372
  11. Cattoni A, Nicolosi ML, Capitoli G, et al. Pubertal attainment and Leydig cell function following pediatric hematopoietic stem cell transplantation: a three-decade longitudinal assessment. Front Endocrinol. 2023;14:1292683. DOI PMID: 38152128
  12. Yin X, Qian L, Dong G, et al. Successful gene therapy for transfusion-dependent α-thalassemia: a case report. Blood Adv. 2026;10(14):5153-5160. DOI PMID: 42200456
  13. Bartelink IH, Lalmohamed A, van Reij EM, et al. Association of busulfan exposure with survival and toxicity after haemopoietic cell transplantation in children and young adults: a multicentre, retrospective cohort analysis. Lancet Haematol. 2016;3(11):e526-e536. DOI PMID: 27746112
  14. Canarutto D, De Gregori S, Recupero S, et al. Busulfan exposure variability and acute toxicity in pediatric patients undergoing hematopoietic stem cell gene therapy. Front Pharmacol. 2026;17:1872108. DOI PMID: 42568941