Between 10% and 15% of sickle cell disease (SCD) are caused by compound heterozygosity for sickle hemoglobin (HbS) and β-thalassemia mutations. Patients with HbS/β⁰-thalassemia usually have microcytic, hypochromic anemia with elevated HbA₂ levels and are symptomatic. Their average HbF level is 8%, although it varies widely among patients. On the other hand, HbS/hereditary persistence of fetal hemoglobin (HPFH) exhibits no anemia or reticulocytosis, and HbF levels persist over 30%. Most of these patients are asymptomatic.¹ Herein, we report three children with the unusual genotype of HbS/Sicilian (δβ)⁰-thalassemia deletion. Two of them suffered from SCD-associated complications in spite of markedly elevated HbF levels with pancellular distribution.

Patient #1: This boy first presented to the hematology clinic at 2 months of age for newborn screening follow-up. His parents are of Asian Indian descent and nonconsanguineous. His father's blood counts are: Hb 13.9 g/dL; mean corpuscular volume (MCV) 68.5 fL; HbA 87.7%; HbA₂ 1.7%; and HbF 10.6%. His mother has sickle cell trait (HbA 61.2%; HbS 35.2%; HbF < 0.8%). The child was examined at ages 5, 11, and again at age 16 months. He is well, has maintained normal growth parameters without splenomegaly or any other SCD-related issues, is on penicillin prophylaxis but not on HU nor has he been transfused. At 11 months’ old, his Hb was 11.9 g/dL, HbS was 48.2%, and HbF was 50.7% (Table S1). His reticulocyte counts were between 1.8% and 2.5%; no other hemolytic markers were examined.

Patient #2: This boy's neonatal hemoglobin screening revealed a Hb FS pattern. His mother is from the Dominican Republic and has sickle cell trait; his father is from Guatemala, and his hemoglobin electrophoresis showed HbA 88%, HbA₂ 2.9%, and HbF 9.1%. This child had a documented pain crisis at age 4 years after swimming in cold water, and another pain episode at age 6. He contracted parvovirus B19 infection and aplastic crisis at age 6 from which he recovered without a packed red blood cells (PRBC) transfusion. At age 8, he had salmonella bacteremia and osteomyelitis of the right femur and was hospitalized. He developed acute chest syndrome during hospitalization and required exchange transfusion. He has fully recovered and continues to do well. His Hb was 11.5 g/dL, HbS 61.9%, and HbF 35.7% (Table S1). He is not on any disease-modifying treatment.

Patient #3: This girl suffered from failure to thrive and frequent pain in her extremities which required analgesics, starting from age 1 year. Her mother is of Spanish descent; her father of Portuguese ancestry. Their laboratory findings are shown in Table S2. At age 8, she had osteomyelitis of her left femur, was hospitalized, and treated with surgical intervention and antibiotics. No micro-organism was cultured. At age 9, she was diagnosed to have HbS/Sicilian (δβ)⁰-thalassemia deletion. She was started on HU at 20 mg/kg/day, followed by clinical improvement. At age 9½, her LD was 280 U/L; total bilirubin was 1.54 mg/dL. At age 11, she continues on HU and is doing well. Her Hb was 13.1 g/dL, HbS was 54.5%, and HbF was 39.3% (Table S1). At age 11½, her LD was 277 U/L; indirect bilirubin was 0.8 mg/dL. Her spleen was modestly enlarged. She has never been transfused.

Results of β-globin gene nucleotide sequencing in these three children were consistent with either homozygosity or hemizygosity for HbS (data not shown). The presence of microcytosis without α-globin gene deletion and with elevated HbF suggested possible β-globin gene deletion. Multiplex ligation-probe amplification (MLPA) revealed a deletion involving δ- and β-globin genes, possibly Sicilian (δβ)⁰-thalassemia deletion (data not shown). A gap-polymerase chain reaction (PCR) test designed specifically for this deletion was done. The results confirmed that all three children were compound heterozygous for HbS and Sicilian (δβ)⁰-thalassemia deletion (Figure 1A,B).

HbS haplotyping was done by genotyping five informative SNPs (Table S3). Patient #1 had the Arab-Indian HbS haplotype consistent with his being of Asian Indian descent; patient #2 had the Benin haplotype; patient #3 had the Bantu (CAR) haplotype. The Sicilian (δβ)⁰-thalassemia deletions found in these three unrelated children had identical haplotype, different from all major HbS haplotypes.

The blood counts, hemoglobin analyses by capillary electrophoresis, genetic analyses of their β-globin genotypes, and F-cells determination by flow cytometric analysis were carried out in patient #3 and her parents (Table S2; Figure 1C). The father is heterozygous for the Sicilian (δβ)⁰-thalassaemia deletion and has heterocellular distribution of HbF, with F-cells accouting for 40.3%. The mother is heterozygous for HbS and has few F-cells accounting for 2.93%. Their daughter (patient #3) is compound heterozygous for HbS/Sicilian (δβ)⁰-thalassemia deletion. She has pancellular distribution of HbF, with F-cells accounting for 92.9%.

Sicilian (δβ)⁰-thalassemia deletion is a deletion of 13.4 kb extending from IVS II of the δ-globin gene to a L1 element 3′ downstream of β-globin gene (Figure 1B).^{2, 3} Adult carriers of this deletion have microcytosis, elevated HbF (9.6 ± 3.5%), and decreased HbA₂.^{3, 4} The fathers of all three patients have HbF of 10.6%, 9.1%, and 8.6%, respectively, consistent with their being heterozygous for the Sicilian (δβ)⁰-thalassemia deletion.

All three patients had markedly elevated HbF. Patient #1 had 50.7% HbF at 11 months of age. He had the Arab-Indian HbS haplotype. The mean HbF level in HbSS patients with the Arab-Indian haplotype is around 18%, much higher than any of the four other major HbS haplotypes found in patients of African descent. Patients #2 and #3 had Benin and Bantu haplotypes, respectively. The mean HbF level in HbSS patients with these haplotypes is in the range between 5% and 7%.

Patients with SCD and elevated HbF usually have mild disease. All three children reported here had HbF levels of 50.7%, 35.7%, and 39.3%, respectively. Patient #1 is well at age 16 months. Additional follow-up is needed to ascertain his disease severity. Patient #2 had history of rare pain episodes and developed salmonella bacteremia, osteomyelitis, and acute chest syndrome at age 8 years. Patient #3 had repeated pain episodes since age 1 and developed osteomyelitis at age 8. In patients #1 and #3, their reticuloctye counts, LD, and bilirubin levels were not elevated as in patients with brisk hemolysis.

Patient #3 responded well to treatment with HU, with resolution of vaso-occlusive pain and improved growth and well-being. Her HbF levels prior to HU treatment were not known but likely elevated as in the other two children. After HU, her HbF levels were not significantly different from patient #2 (Table S1). The clinical beneficial effect of HU is thought to be primarily due to the up-regulation of HbF expression. However, HU can cause decreases in neutrophil and platelet counts, enhance nitric oxide production, downregulate endothelial adhesion molecules, and alter SCD erythrocyte membrane structure.⁵ It is conceivable that one or more of the above might have led to clinical improvement in patient #3.

F-cell distribution was determined by flow cytometry in patient #3 and her parents. Her father is heterozygous for the Sicilian (δβ)⁰-thalassemia deletion and has heterocellular distribution of HbF in his peripheral blood erythrocytes (Figure 1C). In contrast, patient #3 has pancellular distribution of HbF (Figure 1C).

Another girl with HbS/Sicilian (δβ)⁰-thalassemia deletion was reported in 2017 (Patient #4 in Table S1). At 1 year, her HbF was 41.2%. At 18 months of age, she remained well and was not on HU. This is the only case confirmed with DNA-based diagnosis reported in the literature.

Cancio et al. reported nine children who were 4–17 years old with HbS/black (^Aγδβ)⁰-thalassemia.⁶ Their mean Hb was 11.7 ± 1.5; HbF was 26.2 ± 4.3%. All of them suffered from SCD-related complications including acute chest syndrome and avascular necrosis. An 18-year-old African American young man with HbS/black HPFH 2 (Hb 13.9 g/dL; HbF 30.3%) developed massive splenic infarction.⁷ HbSE disease might be comparable to HbS/β⁺-thalassemia. Children and adolescents with HbSE disease are usually well. However, more than half of those aged 20 years and older have SCD-related complications.⁸

HbF is the most important genetic modifier to modulate SCD disease severity, based on its ability to inhibit deoxy-HbS polymerization. It is estimated that 30% or approximately 10 pg of HbF in each HbSS erythrocyte is optimal for the inhibition of deoxy-HbS polymerization. Pancellular distribution of HbF in HbSS erythrocytes is essential to achieve clinical benefit.⁹ Similar to other cases published in the literature, two of the 3 children reported here suffered from SCD complications in spite of markedly elevated HbF levels with pancellular distribution. These clinical observations further illustrate the complexity of SCD pathophysiology.¹⁰ It is conceivable that in patients #2 and #3, while their HbF distribution is pancellular based on flow cytometric analysis, yet HbF present in each erythrocyte is highly variable, some with a lot and others with little which might not be sufficient to negate deoxy-HbS polymerization leading to sickling and related complications. Reliable assay for HbF quantitation in individual erythrocyte is needed to validate this hypothesis.

10% 到 15% 的镰状细胞病 (SCD) 是由镰状血红蛋白 (HbS) 和 β-地中海贫血突变的复合杂合性引起的。^{HbS/β 0}地中海贫血患者通常有小红细胞、低色素性贫血伴 HbA ₂水平升高并且有症状。他们的平均 HbF 水平为 8%，尽管患者之间差异很大。另一方面，HbS/遗传性胎儿血红蛋白持续性 (HPFH) 没有表现出贫血或网织红细胞增多症，并且 HbF 水平持续超过 30%。这些患者中的大多数是无症状的。¹在此，我们报告了三名具有 HbS/西西里 (δβ) 基因型异常的儿童⁰-地中海贫血缺失。尽管全细胞分布的 HbF 水平显着升高，但其中两人患有 SCD 相关并发症。

患者 #1：这个男孩在 2 个月大时首次到血液科诊所进行新生儿筛查随访。他的父母是亚裔印度裔，非近亲。他父亲的血细胞计数是：Hb 13.9 g/dL；平均红细胞体积 (MCV) 68.5 fL；HbA 87.7%；HbA ₂ 1.7%；和 HbF 10.6%。他的母亲具有镰状细胞特征（HbA 61.2%；HbS 35.2%；HbF < 0.8%）。孩子在 5 岁、11 岁和 16 个月时再次接受检查。他很好，保持正常的生长参数，没有脾肿大或任何其他 SCD 相关问题，正在接受青霉素预防，但没有使用 HU，也没有输血。在 11 个月大时，他的 Hb 为 11.9 g/dL，HbS 为 48.2%，HbF 为 50.7%（表 S1）。他的网织红细胞计数在 1.8% 和 2.5% 之间；未检查其他溶血标志物。

患者 #2：这个男孩的新生儿血红蛋白筛查显示 Hb FS 模式。他的母亲来自多米尼加共和国，具有镰状细胞特征；他的父亲是危地马拉人，他的血红蛋白电泳显示HbA 88%，HbA ₂2.9%，HbF 9.1%。这个孩子在 4 岁时在冷水中游泳后出现疼痛危机，并在 6 岁时再次出现疼痛发作。他在 6 岁时感染了细小病毒 B19 和再生障碍危机，在没有输注浓缩红细胞 (PRBC) 的情况下康复. 8 岁时，他患有沙门氏菌菌血症和右侧股骨骨髓炎，并住院治疗。他在住院期间出现了急性胸部综合征，需要换血。他已经完全康复并继续表现良好。他的 Hb 为 11.5 g/dL，HbS 为 61.9%，HbF 为 35.7%（表 S1）。他没有接受任何改变疾病的治疗。

患者#3：这个女孩从 1 岁开始就患有发育迟缓和四肢频繁疼痛，需要镇痛药。她的母亲有西班牙血统；她的父亲是葡萄牙血统。他们的实验室发现如表 S2 所示。8 岁时，她患有左股骨骨髓炎，住院治疗，并接受了手术干预和抗生素治疗。没有培养微生物。9 岁时，她被诊断出患有 HbS/西西里 (δβ) ⁰-地中海贫血缺失。她以 20 mg/kg/天的剂量开始 HU，随后临床改善。在 9½ 岁时，她的 LD 为 280 U/L；总胆红素为 1.54 mg/dL。11 岁时，她继续参​​加 HU 并且表现良好。她的 Hb 为 13.1 g/dL，HbS 为 54.5%，HbF 为 39.3%（表 S1）。在 11½ 岁时，她的 LD 为 277 U/L；间接胆红素为 0.8 mg/dL。她的脾脏适度肿大。她从未输过血。

这三名儿童的 β-珠蛋白基因核苷酸测序结果与 HbS 的纯合子或半合子一致（数据未显示）。没有 α-珠蛋白基因缺失和 HbF 升高的小红细胞症的存在表明可能存在 β-珠蛋白基因缺失。多重连接探针扩增 (MLPA) 揭示了涉及 δ- 和 β-珠蛋白基因的缺失，可能是西西里 (δβ) ⁰ -地中海贫血缺失（数据未显示）。进行了专门为这种缺失设计的间隙聚合酶链反应 (PCR) 测试。结果证实，所有三个儿童都是 HbS 和西西里 (δβ) ^0-地中海贫血缺失的复合杂合子 (图 1A、B)。

通过对五个信息丰富的 SNP 进行基因分型来完成 HbS 单倍型分析（表 S3）。患者 #1 的阿拉伯-印度 HbS 单倍型与他的亚裔印度血统一致；患者#2 具有贝宁单倍型；患者#3 具有 Bantu (CAR) 单倍型。在这三个无关儿童中发现的西西里 (δβ) ⁰ -地中海贫血缺失具有相同的单倍型，不同于所有主要的 HbS 单倍型。

对 3 号患者及其父母进行了血细胞计数、毛细管电泳血红蛋白分析、其 β 珠蛋白基因型的遗传分析以及流式细胞术分析的 F 细胞测定（表 S2；图 1C）。父亲是西西里 (δβ) ⁰ -地贫缺失的杂合子，具有 HbF 的异细胞分布，其中 F 细胞占 40.3%。母亲是 HbS 杂合子，F 细胞很少，占 2.93%。他们的女儿（患者#3）是 HbS/西西里（δβ）^0-地中海贫血缺失的复合杂合子。她具有HbF的全细胞分布，其中F细胞占92.9%。

西西里 (δβ) ^0-地中海贫血缺失是从 δ-珠蛋白基因的 IVS II 延伸到 β-珠蛋白基因下游 3' 的 L1 元件的 13.4 kb 缺失（图 1B）。^{2, 3}这种缺失的成年携带者有小红细胞症、HbF 升高 (9.6 ± 3.5%) 和 HbA ₂降低。^{3, 4}所有三名患者的父亲的 HbF 分别为 10.6%、9.1% 和 8.6%，这与他们是西西里 (δβ) ⁰ -地中海贫血缺失的杂合子一致。

所有三名患者的 HbF 均显着升高。1 号患者在 11 个月大时 HbF 为 50.7%。他有阿拉伯-印度 HbS 单倍型。阿拉伯-印度单倍型 HbSS 患者的平均 HbF 水平约为 18%，远高于在非洲裔患者中发现的其他四种主要 HbS 单倍型中的任何一种。患者#2 和#3 分别具有贝宁和班图单倍型。具有这些单倍型的 HbSS 患者的平均 HbF 水平在 5% 和 7% 之间。

SCD 和 HbF 升高的患者通常病情较轻。这里报告的所有三名儿童的 HbF 水平分别为 50.7%、35.7% 和 39.3%。1 号患者 16 个月大。需要额外的随访以确定他的疾病严重程度。患者 #2 有罕见的疼痛发作史，并在 8 岁时出现沙门氏菌菌血症、骨髓炎和急性胸部综合征。3 号患者从 1 岁开始反复疼痛发作，8 岁时出现骨髓炎。在 1 号和 3 号患者中，他们的网状细胞计数、LD 和胆红素水平没有像快速溶血患者那样升高。

3 号患者对 HU 治疗反应良好，血管闭塞性疼痛得到缓解，生长和幸福感得到改善。她在 HU 治疗前的 HbF 水平未知，但可能与其他两个孩子一样升高。HU 后，她的 HbF 水平与患者 #2 没有显着差异（表 S1）。HU 的临床有益效果被认为主要是由于 HbF 表达的上调。然而，HU 可导致中性粒细胞和血小板计数减少，增加一氧化氮的产生，下调内皮粘附分子，并改变 SCD 红细胞膜结构。⁵可以想象，上述一项或多项可能导致患者#3 的临床改善。

F 细胞分布由患者#3 及其父母的流式细胞术确定。她的父亲是西西里 (δβ) ⁰ -地中海贫血缺失的杂合子，并且在他的外周血红细胞中具有 HbF 的异细胞分布（图 1C）。相比之下，患者 #3 具有 HbF 的全细胞分布（图 1C）。

2017 年报告了另一名患有 HbS/西西里 (δβ) ^0-地中海贫血缺失的女孩（表 S1 中的患者 #4）。1 年时，她的 HbF 为 41.2%。在 18 个月大时，她身体一直很好，没有上 HU。这是文献中报道的唯一通过基于 DNA 的诊断得到证实的病例。

坎西奥等人。报道了 9 名 4-17 岁患有 HbS/黑色 ( ^A γδβ) ^0-地中海贫血的儿童。⁶他们的平均 Hb 为 11.7 ± 1.5；HbF 为 26.2 ± 4.3%。他们都患有与 SCD 相关的并发症，包括急性胸部综合征和缺血性坏死。一名患有 HbS/黑色 HPFH 2（Hb 13.9 g/dL；HbF 30.3%）的 18 岁非洲裔美国年轻人出现大面积脾梗死。⁷ HbSE 疾病可能与 HbS/β ⁺ -地中海贫血相当。患有 HbSE 疾病的儿童和青少年通常情况良好。然而，超过一半的 20 岁及以上的人患有 SCD 相关并发症。⁸

基于 HbF 抑制脱氧 HbS 聚合的能力，HbF 是调节 SCD 疾病严重程度的最重要的遗传修饰剂。据估计，每个 HbSS 红细胞中 30% 或大约 10 pg 的 HbF 对于抑制脱氧-HbS 聚合是最佳的。HbSS 红细胞中 HbF 的全细胞分布对于获得临床益处至关重要。⁹与文献中发表的其他病例相似，尽管全细胞分布的 HbF 水平显着升高，但此处报道的 3 名儿童中有 2 名患有 SCD 并发症。这些临床观察进一步说明了 SCD 病理生理学的复杂性。¹⁰可以想象，在患者 #2 和 #3 中，虽然他们的 HbF 分布是基于流式细胞术分析的全细胞分布，但每个红细胞中存在的 HbF 是高度可变的，一些具有很多而另一些很少，这可能不足以否定脱氧-HbS 聚合导致镰状化和相关并发症。需要对单个红细胞中 HbF 定量进行可靠测定来验证这一假设。