By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Dr Paul Veys, Department of Bone Marrow transplantation, Great Ormond Street Hospital for Children NHS Trust, London WC1N 3JH, UK. E-mail: email@example.com
Umbilical cord blood transplant (UCBT) is associated with impaired early immune reconstitution. This might be explained by a lower T-cell dose infused, the naivety of cord blood T-cells and the use of in vivo T-cell depletion. We studied the pattern of early immune reconstitution and the clinical outcome of children undergoing unrelated UCBT when in vivo T-cell depletion was omitted. Thirty children affected by malignancies (46%) or immunodeficiencies (54%) underwent an unrelated UCBT. Prospective assessment of immune reconstitution and clinical outcome was performed. We observed an unprecedented CD4+ T-cell reconstitution, with a median cell count at 30 and 60 d post UCBT of 0·3 × 109/l and 0·56 × 109/l, respectively. Early T-cell expansion was thymic-independent, with a rapid shift from naïve to central memory phenotype and early regulatory T-cell recovery. Viral infections were frequent (63%) but resolved rapidly in most cases and virus-specific T-lymphocytes were detected within 2 months post-UCBT. Acute graft-versus-host disease (GvHD) was frequent (grade II = 34%, grade III–IV = 16%) but steroid responsive, and the incidence of chronic GvHD was low (14%). The omission of in vivo T-cell depletion promotes a unique thymic-independent CD4+ T-cell reconstitution after unrelated UCBT in children. We postulate that this relates to the specific immunological and ontological qualities of fetal-derived lymphocytes.
Haematopoietic cell transplantation (HCT) is followed by a variable period of immune deficiency, the length and severity of which is dependent on the choice of stem cell source, in vivo and in vitro graft manipulation, treatment of graft-versus-host disease (GvHD) and residual recipient thymopoiesis (Seggewiss & Einsele, 2010).
The pattern of T-cell recovery after transplant generally follows two phases: the antigen-driven clonal expansion of memory T lymphocytes contained in the graft provide early immunity during the first months, while a new wave of post-thymic naïve T-lymphocytes contributes to the long-term immune reconstitution after HCT.
Importantly, the speed and quality of immune recovery after transplant correlates with the relative risk of developing life-threatening opportunistic infections and relapse of malignant disease (Ochs et al, 1995; Parkman et al, 2006).
Compared to other stem cell sources, umbilical cord blood transplantation (UCBT) is characterized by an impairment of early immune reconstitution: recovery of T-cells to normal levels is slow, occurring at 9–12 months post-transplant, generally with the time of thymic recovery (Thomson et al, 2000; Niehues et al, 2001; Komanduri et al, 2007) and infections account for >50% non-relapse mortality in the first 100 d after transplantation (Rubinstein et al, 1998).
It is currently unclear whether this delayed T-cell recovery reflects the quantitative and qualitative characteristics of cord blood compared to recipients of bone marrow transplants (patients who undergo UCBT receive one log less T-cells, all of which are of a naïve phenotype) (Arcese et al, 1998), or the use of in vivo T-cell depletion to prevent GvHD/rejection (Small et al, 1999).
Given that the risk of severe GvHD after UCBT is reduced, even in the context of HLA disparity (Rocha et al, 2004), we hypothesized that the omission of in vivo T-cell depletion may promote early immune reconstitution and reduce infection-related morbidity and mortality in children undergoing unrelated UCBT.
We describe, for the first time, a novel pattern of early immune reconstitution after HCT: cord blood naïve CD4+ T-lymphocytes show an unprecedented thymic-independent peripheral expansion, with a rapid shift to central memory phenotype, broad T-cell receptor (TCR) repertoire and detection of virus-specific response within 2 months after UCBT. This pattern of T-cell reconstitution after UCBT is unique and most likely dependent on the specific immunological and ontological qualities of cord blood T-lymphocytes that are discussed. This translated into low infection-related mortality in a high-risk population of children undergoing unrelated UCBT.
Patients, material and methods
A total of 30 consecutive children undergoing unrelated UCBT within the Blood and Marrow Transplantation Unit at Great Ormond Street Hospital, London, were studied between April 2008 and September 2010.
Children affected by either high risk malignancies or severe primary immunodeficiency diseases were eligible for inclusion in this study if they lacked a 10/10 HLA A-, HLA B-, HLA C-, HLA DRB1- and HLA DQ-matched related or unrelated donor.
The study was approved by the Great Ormond Street Hospital’s Institutional Review Board (protocol number 05/Q0508/61) and written informed consent was obtained from patients’ parents or legal guardians in each case according to the Declaration of Helsinki.
Selection of UCB units
UCB units were obtained from national and international cord blood banks. UCB units were required to be HLA matched to the patient at 4/6 or greater HLA A-, HLA B- and HLA DRB1 antigens as assessed by HLA A, B serological typing and HLA DRB1 high resolution allele typing.
The minimum total nucleated cell (TNC) dose required for the transplant was 3, 4 and 5 × 107/kg in case of a 6/6, 5/6 or 4/6 HLA-matched UCB unit respectively.
Children who did not meet these criteria received a double UCBT: both cord blood units were required to be a 4/6 or greater HLA A, HLA B and HLA DRB1 match with each other and the patient. Prior to transplantation, full allelic typing of HLA A, B, C, DRB1 and DQB1 antigens was performed on the cord sample and from April 2010 units with more than three allelic mismatches were not selected.
Details of conditioning regimen, GvHD prophylaxis and cell dose infused are summarized in Table I. Nineteen patients (63%) received a treosulfan-based preparative regimen, while 10 children affected by malignancies (33%) received a standard myeloablative conditioning regimen. One child affected by severe combined immunodeficiency received an unconditioned graft.
Table I. Clinical features of study subjects.
CMV, cyomegalovirus; TBI, total body irradiation; Bu, Busulfan; GvHD, graft-versus-host disease; HLA, human leucocyte antigen, TNC, total nucleated cells.
*Ciclosporin was given from day −3, adjusted to a trough plasma level of 150–250 ng/ml, and mycophenolate mofetil from day 0 to day +28, then tapered over 3 weeks in the absence of GvHD in 29/30 patients.
Number of patients (male/female)
Median age (years) at transplant (range)
Severe combined immunodeficiency (SCID)
Haemophagocytic lymphohistiocytosis (HLH)
Acute lymphoblastic leukaemia
Acute myeloid leukaemia
Chronic myeloid leukaemia
Pre-transplant recipient’s CMV seropositivity
Pre-transplant viral diseases
Pre-transplant fungal diseases
Standard myeloablative conditioning (TBI or Bu-based)
Degree of HLA match (n = 34, since four double cords)
4/6, 5/6, 6/6
3/34, 15/34, 16/34
6/10, 7/10, 8/10, 9/10, 10/10
2/34, 7/34, 8/34, 14/34, 3/34
TNC × 107/kg, median (range)
CD34+ cell × 105/kg, median (range)
CD3+ cell × 106/kg, median (range)
GvHD prophylaxis consisted of ciclosporin (CsA, from day −3, adjusted to a trough plasma level of 150–250 ng/ml) and mycophenolate mofetil (MMF, from day 0 to day +28, then tapered over 3 weeks in the absence of GvHD) in 29/30 patients, with additional methylprednisolone in three cases. No patient received either anti-thymocyte globulin (ATG) or Alemtuzumab. The majority of the cord blood units (54%) were HLA-mismatched with the recipient. Four children (13%) received a double UCBT in order to reach the minimum TNC dose required.
Supportive therapy consisted of ciprofloxacin (from day −10, until neutrophil count ≥ 1 × 109/l), acyclovir (from day −10, until 1 year post-transplant) and itraconazole (from day −11, until neutrophil count ≥ 1 × 109/l and no steroid treatment).
Prevention of Pneumocystis jiroveci pneumonia included cotrimoxazole daily from day −10 to day −1 and then, after myeloid recovery, until a CD4+ T-cell count >0·3 × 109/l and absence of chronic GvHD/immunosuppressive treatment.
Patients received recombinant human granulocyte colony-stimulating factor (G-CSF) at 5 μg/kg per day from day +14, until an absolute neutrophil count >1 × 109/l was reached for two consecutive days.
From August 2010 children also received antibiotic prophylaxis with vancomycin (400 mg/m2 from day +1, until neutrophil count >0·2 × 109/l), due to a high incidence of sepsis with Streptococcus viridans (Kurt et al, 2008).
From January 2010 children were commenced on gut rest and received total parenteral nutrition (TPN) from day −10 until myeloid engraftment.
The diagnosis of acute GvHD was made clinically, and confirmed pathologically with skin, mucosal or liver biopsy whenever possible. Grading of acute GvHD was performed according to the Seattle criteria (Glucksberg et al, 1974). Chronic GvHD was assessed and scored according to the National Institute of Health (NIH) criteria (Filipovich et al, 2005).
Analysed of factors which predicted early immune reconstitution was performed using spss® 18.0 (IBM Corporation, New York, NY, USA). Dependent variables of CD3+ and CD3+CD4+, lymphocyte counts at 1, 2 and 3 months post-transplant were normalized by loge transformation. To reduce the impact of variability of individual lymphocyte subset measurements, mean lymphocyte subset counts for 1–3 months were also analysed.
Predictive factors of age (days at transplant), CD34+ and CD3+ cord cell dose (all loge transformed), underlying disease, conditioning regimen, HLA mismatch and number of cord units infused were assessed for their influence on immune reconstitution. Categorical factors were grouped as follows; underlying disease (malignant versus non-malignant), conditioning (reduced intensity versus full myeloablative), 10-locus HLA mismatch (10/10 vs. 9 & 8/10 vs. 7 & 6/10).
Univariate associations between each predictor and each lymphocyte subset reading were analysed using Pearson’s correlation test, t-test or one-way analysis of variance (anova) as appropriate. Predictors with a P value of <0·2 were then assessed in multivariate analyses of covariance (ancova), and retained if they reached a significance of P < 0·05.
Additional material and methods
Details on monitoring of cytomegalovirus (CMV), adenovirus and Epstein–Barr virus (EBV) infection, engraftment, donor chimerism and lymphocyte subsets, detection of virus specific responses, perforin expression, Tregs/measurement of FOXP3 expression, T-cell receptor excision circles (TREC) analysis, spectratype analysis and statistical analysis are described in the Supporting Information.
Patients and clinical outcome
Thirty consecutive children affected by either high risk or relapsed/refractory malignancies (46%) or primary immunodeficiency diseases (54%) underwent unrelated UCBT during the study period. Immune reconstitution was assessed over a median follow-up period of 12 months (range, 2–32). A notable proportion of the patients had pre-transplant co-morbidities including viral (27%) or fungal (17%) infections. All evaluable patients (29/30) experienced donor engraftment; one child died prior to engraftment (day +2) due to conditioning-related toxicity. Myeloid recovery occurred at a median time of 22 d (range, 13–38), while platelet recovery was recorded at a median time of 42 d (range, 17–123) post-UCBT. At last follow-up, 25/29 patients remained 100% donor in all cell lineages, while four patients exhibited mixed donor chimerism in the myeloid compartment only (range: 0–62% donor engraftment in CD15+ cells): one of these children did not receive conditioning, whilst three children received a treosulfan-based preparative regimen. Notably, all evaluable patients showed 100% donor engraftment in the T-cell compartment.
Eighteen patients experienced a peri-engraftment syndrome (Patel et al, 2010) within 14 d of transplantation and responded rapidly to a short course of methylprednisolone (1–2 mg/kg per d). Interestingly, since the introduction of complete gut rest from the start of conditioning until myeloid engraftment, we have observed a reduced incidence of peri-engraftment syndrome (15/17 episodes without gut rest versus 3/13 episodes with gut rest).
Grade II and grade III–IV acute GvHD occurred in 10 (34%) and 5 (16%) patients respectively and was generally steroid-responsive. Of the latter group, two had undergone 6/10 HLA-matched UCBT for refractory haemophagocytic lymphohistiocytosis (HLH), one received a 7/10 HLA-matched graft, one a 9/10 graft and one a 10/10 HLA-matched graft.
Four children (14%) developed chronic GvHD affecting the skin (n = 3) or the gastrointestinal tract (n = 1). Chronic skin GvHD was scored as mild in three cases and in one case of vitiligo it resolved completely; one child was affected by moderate chronic gut GvHD and required immunosuppression at the last follow-up.
Leukaemia relapse occurred in 5/14 patients (36%): four died of disease recurrence, while one underwent a second transplant and was still alive and in complete remission at last follow-up. At last follow-up 23 children were alive (76%).
Bacterial sepsis, occurring mostly during the neutropenic phase, was frequent (n =19 patients, n =32 episodes of bacteraemia). Interestingly, seven children developed Streptococcus viridans sepsis, therefore vancomycin prophylaxis was started from August 2010.
Viral infections, mostly within the first 2 months post-UCBT, were also frequent (63% of the patients) and included: adenoviraemia (n =4, 14%), CMV viraemia (n =4, 14%), parainfluenza 3 (n =3, 10%) and respiratory syncytial virus (RSV, n =3, 10%) respiratory tract infections, EBV viraemia with no lymphoproliferative disease (n =3, 10%), human herpes virus 6 (HHV6) encephalitis (n =1, 3%) and varicella zoster virus (VZV) infection (n =1, 3%). One child requiring high dose steroids for refractory HLH at the time of the transplant died of parainfluenza 3 pneumonia 6 weeks post-UCBT, while another child died 9 months after transplant from acute respiratory distress syndrome/pulmonary hypertension and RSV infection in the context of late acute GvHD. All remaining patients recovered fully from viral infections. Overall, the infection-related mortality was low (2/30 = 7%).
Reconstitution of CD4+ and CD8+ T-cells after UCBT
The kinetics of T-cell recovery after UCBT are shown in Fig 1. We observed a rapid increase of T-cell counts over the first 2 months after transplant, especially of the CD4+ T-cell compartment, with a median CD4+ T-cell count at 30 and 60 d after UCBT of 0·31 × 109/l (range, 0·06–1·89) and 0·56 × 109/l (range, 0·28–1·69), respectively. CD8+ T-cells recovered more slowly than CD4+ T-cells, with a CD4:CD8 ratio at 1, 2, 6 and 12 months post-UCBT of 5, 3·5, 2 and 2·5, respectively.
At last follow-up, all 23 patients had normal CD4+ T-cell counts for their age (Comans-Bitter et al, 1997), with a median time to CD4+ T-cell recovery of 2 months (range, 1–12), while 18 patients achieved a normal CD8+ T-cell count at a median time of 3 months post-UCBT (range, 1–12). Younger patients demonstrated more rapid CD3+ and CD4+ T-cell reconstitution, but neither CD34+ nor CD3+ cord cell dose were correlated with recovery of any lymphocyte subset at the time intervals measured (Table SI).
Shift from naïve to memory phenotype
The percentage of CD4+CDRA+CD27+ naïve T-cells decreased within the first weeks post-UCBT (Fig 2A), with a very rapid shift to the CD4+CDRA−CD27+ central memory T-cell phenotype. Naïve T-cells numerically rose again 1 year after transplantation, most likely due to the return of thymopoiesis.
Relationship between CD4+ T-cell recovery and of TREC numbers
Thymic output was monitored early after UCBT in 17 children: the correlation between CD4+ T-cell recovery and TREC numbers at 1, 2, 3, 6 and 12 months after UCBT is shown in Fig 2B. During the first 6 months after transplant high CD4+ T-cell counts were associated with low TREC numbers, suggesting that the early T-cell recovery was independent of thymic activity and was due to the homeostatic peripheral expansion of cord blood lymphocytes infused with the graft. Nevertheless, 12 months after UCBT we observed a concurrent increase in CD4+ T-cells and TREC numbers, suggesting that thymopoiesis contributed to late T-cell reconstitution.
T-cell repertoire diversity
We monitored the T-cell repertoire of 20 children studied at 1, 2, 3, 6 and 12 months after UCBT by TCR spectratyping. Surprisingly most VB families were polyclonal, with a normal spectratype distributed in a Gaussian fashion as early as 30 d after UCBT (example in Fig 3A).
We observed a median of 6 and 7 different TCR CDR3 size classes per VB family 30 and 60 d after UCBT respectively (data not shown). The VB density peak histograms were also analysed by a web-based automated program for the Kullback–Leibler divergence (DKL) (Douek et al, 1998): the higher the DKL, the greater the divergence from a normal profile and the more oligoclonal the distribution of peaks (normal DKL value < 0·1). The longitudinal monitoring of T-cell repertoire diversity is represented in Fig 3B: the DKL score was low early after UCBT, suggesting that the T-cell repertoire was polyclonal in most patients early after the transplant.
Functional T-cell responses
Peripheral blood mononuclear cells from 16 patients were analysed at 2, 3 and 4–6 months after transplant for their ability to secrete cytokines after stimulation with the non-specific mitogen [phytohaemagglutinin (PHA)] and for antigen-specific responses to adenoviral and CMV antigens using γ-interferon (IFNγ)-enzyme-linked immunosorbent spot (ELISPOT).
The majority of evaluable patients (11/16, 70%) were able to secrete IFNγ in response to PHA stimulation at these early time points (Fig 4A). Three children showed CMV-specific response at 2–4 months after UCBT; two had experienced previous CMV viraemia. Adenovirus-specific CD4+ T-cells were detected using an INFγ secretion and capture assay 2 months post-transplant in one child with adenovirus viraemia (Fig 4B). These data indicate that the T-cells recovering at early time-points post-UCBT are functional.
Regulatory T cells
Blood samples collected for this study were of small volume and were cryopreserved for non-routine analysis, precluding the assessment of regulatory T cells (Treg) by conventional intracellular staining for FOXP3. Instead we adopted the CD4+CD25+CD127low cell surface marking as a surrogate phenotype for Treg quantification (Liu et al, 2006). The median percentage of CD4+CD25+CD127dim cells at 1 and 2 months post UCBT in 14 evaluable patients was 7·4% (range, 0–70·7%) and 8·1% (range, 0·2–24%) respectively, indicating the early presence of cells with regulatory potential. Further evidence of Treg reconstitution was derived from the measurement of FOXP3 mRNA by quantitative polymerase chain reaction, which confirmed active expression of this transcription factor in peripheral blood cells, albeit at variable levels; within 3 months post-UCBT the levels of FOXP3 in peripheral blood mononuclear cells of 14 children ranged between 5·5% and 1422% (median 37%) of expression levels found in healthy children.
These data suggest that while early CD4+ T-cell recovery after UCBT is not primarily due to Tregs, this population is nonetheless detectable at early time-points after UCBT.
Reconstitution of natural killer (NK) and B cells and IgM levels
The pattern of NK and B-cell recovery after UCBT is shown in Fig 5. NK recovery was fast: 29 children achieved normal counts for their age at a median time of 1 month post-UCBT (range, 1–3 months).
B-cell reconstitution was also prompt: 22/26 evaluable children achieved normal B-cell counts for their age, at a median time of 2 months post-UCBT (range, 1–12). Twenty children achieved normal levels of IgM, at a median time of 4 months (range, 1–12).
Several reports suggest that UCBT is associated with impaired immune reconstitution because of delayed T-cell recovery (Komanduri et al, 2007), lack of transfer of antigen-experienced lymphocytes, deficient cytokine production (Lewis et al, 1991) and high incidence of early infection-related mortality (accounting for up to 50% of transplant-related deaths in a large study reported by Rubinstein et al (1998).
The significant lower cell dose infused with a cord blood unit compared to bone marrow or peripheral blood and the use of in vivo T-cell depletion (ATG or alemtuzumab) within the conditioning regimen might further explain the prolonged lymphopenia after UCBT.
Our results show for the first time a unique pattern of early immune reconstitution in children undergoing unrelated UCBT when in vivo T-cell depletion is omitted: within 2 months after transplant cord blood lymphocytes undergo rapid thymic-independent peripheral expansion, with a shift from naïve to central memory phenotype, broad TCR repertoire and early detection of virus-specific T-cell response.
We observed an unprecedented early recovery of CD4+ T-cells, with a median cell count of almost 0·6 × 109 cells/l at 2 months after UCBT. At the same time-point, 23 children achieved normal CD4+ T-cell counts for their age. This contrasts markedly to what has been previously published in other paediatric studies of UCBT incorporating in vivo T-cell depletion, where the CD4+ T-cell reconstitution occurred much later. In one study of 27 patients, the median CD4+ T-cell count was 0·15×109 cells/l at 2 months post-UCBT (Thomson et al, 2000) and another study reported only four of 12 children with a CD4+ T-cell count above 0·2×109/l at the same time-point (Moretta et al, 2001). Eurocord analysis of immune reconstitution in a cohort of 63 children undergoing mostly unrelated UCBT with serotherapy demonstrated that CD4+ T-cell recovery occurred at a median of 12 months after transplant (Niehues et al, 2001). More recently, Rénard et al (2011) reported their results on immune reconstitution after unrelated UCBT in 112 children: the median time to achieve a CD4+ T-cell count of 0·2×109/l and 0·5×109/l was 5 and 9 months, respectively. These studies contrast markedly to the rapid kinetics of CD4+ T-cell recovery observed in our cohort of children not treated with in vivo T-cell depletion.
Interestingly, Sauter et al (2011) recently reported their results on immune reconstitution after double unrelated UCBT with no use of ATG in older patients (median age 36 years): in this series the median time to reach a CD4+ T-cell count >0·2×109/l was 4 months.
The older age of the patients, the lower TNC dose infused and the immunological competition between the two cord blood units might not favour an equally rapid T-cell reconstitution.
To our knowledge, these are the best results published so far in terms of early T-cell reconstitution in children after unrelated UCBT, or indeed after any stem cell transplant procedure. A comparison with a previous cohort of 50 children treated at our institution who underwent a HCT from a matched sibling donor without serotherapy (data not shown) showed significantly slower CD4+ T-cell recovery than patients in this study receiving UCBT with no in vivo T-cell depletion (median CD4 T-cell count at 1, 2 and 3 months after matched sibling HCT = 0·19, 0·21 and 0·23 × 109/l, respectively).
Our results also showed that CD8+ T-cell reconstitution was somewhat delayed compared to CD4+ T-cell recovery, with a high CD4:CD8 ratio over the first 6 months after UCBT. Thus we did not observe the characteristic inversion of the CD4:CD8 ratio described after transplant from other stem cell sources (Kook et al, 1997; Lamb et al, 1998). Nevertheless CD8+ T-cell recovery was significantly faster compared to other reports in the UCBT setting (Rénard et al, 2011), with a median time to reach a CD8+ T-cell count above 0·25 × 109/l of 2 months.
Our population of patients might not be entirely representative of those reported in other studies, because the median age of the patients was low, the TNC dose was high, and the majority of children was affected by immunodeficiency disorders and received a well-matched UCB unit. However, when these factors were analysed, we found that only young age was associated with a more rapid CD3+ T-cell recovery, but the speed of lymphocyte reconstitution was not thymic-dependent and it was not found to be associated with the CD3+ T-cell dose infused with the cord blood unit. This is not surprising, as the median age of our patients was low (12 months) and the median lymphocyte dose infused with the graft was high. Moreover, other factors, such as type of disease (malignant versus non-malignant), HLA-matching or conditioning regimen were not shown to be significant in the speed of immune reconstitution.
Moreover, it is generally accepted that delayed T-cell recovery after transplant depends on the age-related decrease in thymic function and CD4+ T-cell recovery normally occurs with return of thymopoiesis (Crooks et al, 2006; Komanduri et al, 2007). Here, we showed that the remarkable CD4+ T-cell reconstitution observed in our study was mediated by a peripheral expansion of the lymphocytes infused with the graft within the first 2 months after UCBT and not by the return of thymopoiesis. In fact, the numbers of peripheral recent thymic emigrants (TRECs) were low for the first 6 months post-UCBT and TRECs only increased between 6 and 12 months after transplant. With the return of thymopoiesis we observed a concomitant further raise in CD4+ T-cell numbers, consistent with the appearance of a new wave of post-thymic T lymphocytes.
Cord blood T lymphocytes are typically CD45RA+ and antigen-inexperienced and major concerns have been raised regarding their ability to provide an efficient immune response after UCBT (Brown & Boussiotis, 2008). Interestingly, we found a very rapid conversion from naïve to central memory phenotype within the first weeks after transplantation. Early and Reen (1999) previously showed that newborn naïve T cells can transform more quickly than their adult counterparts into functionally equivalent central memory T-cells in vitro, a process that may be important to counteract the immature immune environment typical of the newborn. Our findings suggest that cord blood naïve T lymphocytes can be rapidly responsive in vivo to allogeneic/viral stimuli and adopt a memory/effector phenotype early after the transplant.
Another interesting finding was that the monitoring of TCR diversity showed a broad TCR repertoire early after UCBT, with a polyclonal profile of VB families within the first 2 months after transplant. This is very different from what has been described previously (Talvensaari et al, 2002), with TCR repertoires being highly abnormal for up to a year post-UCBT and only becoming more diverse with return of thymopoiesis. Given that cord blood naïve T lymphocytes have a completely formed TCR repertoire that reflect the lack of prior antigen stimulation (Garderet et al, 1998), we postulate that the omission of in vivo T-cell depletion might foster the expansion of lymphocytes with a polyclonal TCR repertoire, therefore promoting a more complete reconstitution of the T-cell compartment.
Fetal naïve CD4+ T-cells show a much higher proliferative response to stimulation with allogeneic cells compared to adult T-cells and appear more likely to adopt a regulatory T-cell fate. We encountered evidence of both these qualities in CD4+ T-cell expansion post-UCBT. This might be due to ontological qualities of cord blood T-cells, such as their origin from fetal stem cells rather than adult stem cells (both of which are present in cord blood) (Mold et al, 2010). Importantly, naïve T-lymphocytes were able to differentiate into antigen-specific T lymphocytes with specificity for viral pathogens (CMV and adenovirus) within 2 months post-UCBT. While viral-specific T-cell responses were not seen in the absence of viral reactivation, four of six evaluable patients with CMV/adenoviraemia showed detectable viral-specific IFNγ responses. The virus-specific response was not hampered by the concurrent treatment of acute GvHD. These findings confirm and extend the results of Cohen et al (2006) who demonstrated the presence of T lymphocytes with specificity for herpes viruses as early as 29 d after UCBT. The rapid T-cell recovery and the detection of donor-derived antigen-specific T-cells is of paramount importance, because infections are a major cause of treatment failure after UCBT (Rubinstein et al, 1998). Moreover, it has been shown that antigen-specific immune reconstitution after UCBT results in decreased leukaemic relapse and improved overall survival (Parkman et al, 2006).
Similarly to previous reports (Thomson et al, 2000; Niehues et al, 2001; Rénard et al, 2011) NK recovery was prompt, with a median time to achieve normal counts of 1 month post-UCBT. B cell recovery was faster than previously described, with a median time to reach normal counts of 2 months after transplant, suggesting that early CD4+ T-cell recovery might help to promote B-cell maturation and Ig class switching.
This functional immune recovery in vitro was paralleled by clinical outcomes in vivo. Although the majority of the children (>50%) experienced a viral or fungal infection after UCBT, the overall mortality due to infectious complications was low (7%).
Historically, serotherapy has been used in conditioning regimens for UCBT both to promote engraftment and reduce GvHD. In our series engraftment was seen in all 29 evaluable patients, with full donor chimerism in 25/29, indicating that serotherapy is not needed for engraftment, perhaps reflecting enhanced graft-versus-marrow alloreactivity.
Conversely, one potential concern if serotherapy is omitted is the increased risk of GvHD. Prior reports indicated a 14–52% incidence of grade II–IV acute GvHD after unrelated UCBT (Rubinstein et al, 1998; Rocha et al, 2004). While we observed a high incidence of grade II–IV acute GvHD (50%), the incidence of severe (≥grade III) acute GvHD (16%) was comparable to those previously described (Rubinstein et al, 1998; Alsultan et al, 2011). As reported in other series (MacMillan et al, 2009; Alsultan et al, 2011), acute GvHD was generally steroid responsive and, most importantly, the pattern and quality of early immune reconstitution was not hampered by steroid treatment.
Given that we infused a T-replete cord unit, we studied a possible relationship between cell dose and incidence of GvHD, but no such correlation was found: this might reflect the young median age of the treated children and subsequent high cell dose infused with the graft. Similarly, the risk of GvHD in this series did not seem to be associated with the degree of HLA mismatch. This may be due to the relatively small number of patients, as it should be noted that both patients undergoing a 6/10 mismatched UCBT developed grade IV acute GvHD. The importance of allele level matching at HLA-A and B and/or whether HLA-C should be considered in the selection of UCB units for transplantation is yet to be determined.
Likewise, the incidence of chronic GvHD was lower than previously reported (14%) (Rubinstein et al, 1998). This might be due to the young age of the recipients, but may also reflect the early expansion of CD4+ regulatory T-cells we observed: several studies both in the animal model and in the clinical setting have shown that CD4+CD25+FOXP3+ T-cells suppress acute and chronic GvHD (Taylor et al, 2002; Zhao et al, 2008). It was recently demonstrated that patients with a lower proportion of CD4+ T-cells with a regulatory phenotype after HCT were at higher risk of developing both acute and chronic GvHD (Zorn et al, 2005; Rezvani et al, 2006). However it should be noted that we adopted a surrogate phenotype of regulatory T-cell quantification (Liu et al, 2006) and further studies with intracytoplasmic staining for FOXP3 will be necessary to confirm this.
In summary, our study shows for the first time that the omission of in vivo T-cell depletion in children undergoing unrelated UCBT promotes a unique pattern of early immune reconstitution with rapid thymic-independent CD4+ T-cell expansion, immunophenotype switch, broad TCR repertoire and virus-specific T-cell response.
Given that this was achieved without compromising engraftment and with acceptable rates of GvHD, we believe that this approach can safely be used to accelerate immune reconstitution after UCBT in children.
This strategy may improve outcomes after UCBT and influence the choice of stem cell donor and, contrary to currently held beliefs, may be particularly useful in children with pre-transplant infectious co-morbidities.
The authors would like to thank the staff of the Immunology laboratory at Great Ormond Street Hospital for their help with sample analysis. This work was supported in part by the National Institute for Health Research Biomedical Research Centres funding scheme.
R.C. performed research, analysed data and wrote the paper, K.G., W.Q. and S.A. performed research and analysed the data, A.J.J.W performed the statistical analysis and wrote the paper, H.Z., C.A.M. and S.D. performed research and analysed the data, C.C. analysed the data, K.R. analysed data and wrote the paper, P.H. performed research and analysed the data, R.H analysed data and wrote the paper, A.S. wrote the paper, C.S.F. and N.G. analysed data and wrote the paper, P.J.A. and P.V. performed research, analysed data and wrote the paper.