Partially corrected X-linked severe combined immunodeficiency: long-term problems and treatment options
- First Online:
- Cite this article as:
- De Ravin, S.S. & Malech, H.L. Immunol Res (2009) 43: 223. doi:10.1007/s12026-008-8073-6
- 204 Views
Rapid progress has been made from the identification of the molecular defects causing X-linked severe combined immune deficiency (X-SCID) to the development of cutting-edge therapeutic approaches such as hematopoietic stem cell transplant and gene therapy for XSCID. Successful treatment of XSCID has created a new population of patients, many of whom are now adolescents and young adults and are facing a variety of chronic problems secondary to partial correction of their underlying disease. This review focuses on the clinical challenges facing these patients (and their caregivers) and provides an overview of some of the treatment options available, including gene therapy.
KeywordsSevere combined immunodeficiencyHematopoietic stem cell transplantGene therapyHaploidenticalT cell recombination excision circlesT cell receptor Vbeta spectratype
Severe combined immunodeficiency (SCID) was first described in 1950  and is characterized clinically by profound defects in humoral and cellular immunity . Mutations leading to SCID occur in more than 12 known genes, some earlier ones identified include adenosine deaminase, the common cytokine receptor gamma chain (γc), Janus kinase 3 (JAK3), IL-7 receptor alpha chain, recombinase activation genes 1 and 2, Artemes, and CD45. The study of these diseases has contributed substantially to our understanding of the human immune system. Recent progress in molecular immunology has provided further characterization of the immune phenotype and the clinical syndrome of SCID (see reviews [3, 4]).
X-linked SCID is the most frequently diagnosed form of SCID, and was coined to mutations in the IL-2 receptor gamma chain in 1993, subsequently renamed the common cytokine receptor gamma chain (γc) [5, 6]. The classical presentation for XSCID occurs in the first few months of life with persistent infections due to opportunistic microbes, persistent diarrhea, and failure to thrive . Causative organisms include Candida albicans, varicella, adenovirus, respiratory syncytial virus, cytomegalovirus, Epstein–Barr virus, and parainfluenza virus. Pneumocystis carinii infection results in a distinct diffuse pneumonitis that often alerts clinicians to an underlying immune deficiency, as does infection by the normally avirulent bacillus Calmette-Guerin (BCG) vaccine [7, 8]. Transplacentally acquired maternal cells may persist in these patients and cause a rash due to GVHD of the skin , or occasionally the liver . As protective maternal antibodies wane after the first few months of life, affected children experience an increased frequency of infection. Diagnosis of SCID is complicated by the fact that multiple infections of the upper respiratory tract and gastrointestinal system are not uncommon in the general population. Thus, presentation with infection alone is not diagnostic and the most frequently performed routine screening test (CBC, complete blood count) seldom includes typing of lymphocyte subsets, a potential diagnostic test. Delayed recognition of SCID may lead to death, long-term organ damage, as well as less favorable outcomes of subsequent interventions , thus highlighting the potential value of neonatal screening. Prior to matched sibling bone marrow transplant in 1968 , patients with XSCID almost always died within the first year of life [14, 15]. The exceptions are of particular interest as they have contributed to the understanding of the molecular pathophysiology of the disease. Mutations resulting in production of low levels of γc protein can complicate bone marrow transplants (BMT) yet also confer some protection that allows survival beyond early childhood [16–19]. Revertant mutations that result in composite phenotypes have also been described , as has one that leads to an Omenn like syndrome .
The common cytokine receptor gamma chain (γc) that is defective in XSCID is a critical component of many cytokine receptors including those for IL-2, -4, -7, -9, -15, and -21. The profound immune dysfunction seen in XSCID underscores the importance of these cytokines for normal cellular growth and differentiation. Activation of γc, an integral membrane protein, normally results in stimulation of a downstream signaling pathway via the Janus kinases (Jaks) and the signal transducers and activators of transcription (STATs) to act upon target genes (reviewed by Leonard [22, 23]). The classic immunophenotype of patients with XSCID is an almost total absence of T lymphocytes and NK cells. The number of B cells is generally normal (or may be elevated in some patients) but their function is abnormal. Interestingly, a canine form of SCID results from γc mutation and shares many phenotypic similarities with human XSCID [24, 25]. In contrast, γc knockout mice display an absence of B cells in addition to the lack of T and NK cells .
Such profound impairment of immunity is life threatening. Early attempts at introducing functional lymphoid cells by allogeneic transplants resulted in fatal graft-versus-host disease (GvHD) . With the advent of HLA typing  and the recognition of the deleterious effects of tissue mismatch in mouse transplant models, the first successful HLA-matched BMT from a sibling resulted in immune reconstitution reported in 1968 . Unfortunately, HLA-matched relatives are often not available as donors [29–32]. During the 1970s, techniques were developed to deplete the immune-competent donor T cells from donor marrow that had previously caused the GvHD . This method increased the success rate of allogeneic HLA-mismatched transplants from related donors, usually parents [33–36, 31]. While more than 90% of SCID patients who receive HLA-identical BMT survive , the overall reported survival rate for haploidentical BMT ranges between 45% and 78% survival [14, 15], although there are significant differences reported from different centers.
While the depletion of donor T cells is generally agreed upon, the utility and toxicity of pre-transplant myeloconditioning chemotherapy for the purpose of increasing donor B cell and NK cell engraftment remains controversial . The absence of T cells and NK cells in XSCID patients suggests an inability to reject the graft, thus obviating the need for pretransplant chemotherapy. Some early reports described no significant advantages in engraftment rate, B cell engraftment and reconstitution, or efficiency of T cell reconstitution in patients who received myeloconditioning compared with those who do not [39, 40]. The outcome of T cell-depleted, haploidentical BMT appeared to correlate with age (infants <3 months do significantly better)  and with the health status of the patient (i.e., patients with infections or organ damage did worse in terms of survival and T or B cell reconstitution). As with many rare diseases, statistical evaluation of different approaches is limited by the numbers of XSCID patients undergoing different treatments at any one center. The largest single center experience with XSCID patients (Duke University Medical Center, NC) reported a survival rate of about 80% (89/112) following BMT without myeloconditioning for SCID in general, with follow-up from 3 months to 18.8 years after transplant. Of the 89 surviving patients, only 26 are alive for 10 year or more .
Some clinical groups have included pretransplantation myeloconditioning in an effort to increase the engraftment rate and to improve T and B cell reconstitution, though at the cost of increased early toxicity , particularly in patients with infection at the time of transplant. In a survey of clinical results in 1990, Fischer et al.  reported that long-term survival was 52% in patients receiving T cell-depleted, HLA-nonidentical BMT from a related donor. The use of a conditioning regimen (CR) including Busulfan led to a higher engraftment rate but was not associated with any significant improvement in survival. However, a recent long-term follow-up with patients following T cell depleted, haploidentical BMT, however, have correlated overall reconstituted B cell immunity and stable naïve T cell numbers with persistent detection of donor CD34s in patients’ marrow, which were absent in patients who did not receive myeloconditioning . In addition to persistent B cell or humoral dysfunction , other notable problems that have developed following a T cell depleted, haploidentical BMT include progressive graft loss or failure, worsening lymphopenia , restricted T cell repertoire, and minimal thymic output as indicated by T cell receptor excision circles .
Poor B or T cell reconstitution may lead to further attempts at BMT, usually following haploidentical BMTs . At one center, up to 21% of 112 patients with SCID received boosters under the same conditions as initial BMT, i.e., T cell depleted and without myeloconditioning . As in the case of initial transplants, myeloconditioning drugs are not uniformly employed at different centers.
Matched unrelated donor transplants
More recently, with significant advances in molecular typing, matched unrelated donor (MUD) BMT have been performed using either a donor marrow program or cord blood , with some reports of superior outcome compared to related, haploidentical BMT . The early problems with the availability of matched unrelated donors have been partially ameliorated by the advent of the National Marrow Donor Program. Furthermore, recent improvements in immunomodulatory agents and growth stimulating factors have also significantly reduced GvHD, and related side effects such as mucositis.
Depending on the transplant center, as well as degree of HLA matching, the outcome following haploidentical donor compared with MUD transplant may be similar or perhaps improved in the MUD group. A two-center study in Canada and Italy reported survival rates of 80.5% in MUD, compared with 52.5% for related haploidentical transplant . Of note, more than 80% of the MUD-transplanted patients in this report demonstrated robust immune reconstitution (including B cell function) compared with a third of haplo-BMT requiring a second transplant, and 39% with indications of T cell abnormalities . Close long-term follow-up of MUD recipients is necessary to make appropriate long-term comparisons, especially when taking into consideration toxicities or mutagenic risks that are potentially associated with chemotherapeutic agents and/or irradiation.
With haploidentical transplants becoming standard treatment over the past 15–20 years, reports of 3 month or even 5-year survival still do not reflect the longer-term sequelae. Over 1,200 immunodeficiency patients have been treated by allogeneic BMT [15, 47]. However, there is scant published literature on the long-term clinical problems and status of XSCID patients after transplant of a T cell depleted, haploidentical BMT. A significant minority of these partially corrected patients have ongoing problems due to inadequate or failing T cell grafts, B cell dysfunction, and NK-cell dysfunction, as well as organ damage from infections before and after transplant. This factor has been a major impetus in the development of current experimental protocols of gene therapy or MUD transplants for XSCID. Here, we review from our NIH experience and from the literature, the type of long-term medical problems affecting the subset of haploidentical-donor transplanted XSCID patients with poor or waning immunity during their second decade of life.
T cell compartment
Functional reconstitution of the T lymphocyte compartment can be evaluated in a number of ways now in addition to simple cell counts, percentage of donor chimerism, and lymphocyte proliferation response to stimulation. In general, T cell immunity is restored 6–12 months after transplant, and a high percentage of transplanted XSCID patients demonstrate lymphocyte counts within normal ranges. Also coinciding with the appearance of naïve T cells following transplant is growth of the thymus, an essentially vestigial organ in XSCID patients. Within the thymus, T cell precursors undergo TCR gene rearrangements by the junction of V(D)J gene segments and by the addition of nucleotides. The process of TCR rearrangement generates DNA episomes or T cell Recombination Excision Circles (TRECs) that can be detected only in newly thymus-derived T cells but not in T cells that develop extrathymically. Since TRECs do not replicate during mitosis, they are diluted at each cell division and therefore TREC content of T cells is an excellent indicator of thymic output. In normal humans, thymic output and TREC content diminish with age. In XSCID patients before transplant, TRECs are undetectable. However, within 3–6 weeks following haploidentical donor transplant of XSCID patients, TRECs begin to appear indicating the emergence of thymic-processed new lymphocytes. While it is encouraging that the progenitor cells can undergo intrathymic differentiation with TRECS rising for the first 1–2 years after transplant of XSCID patients, one study has reported that within 12–14 years, TRECs decline to levels that are only a few percent of normal for their age group, a number comparable to the low level of TRECs seen in normal 80 year olds .
Thymic output can also be assessed ex vivo by surface immunophenotyping of CD45RA+ naïve T cells in the peripheral blood. However, these cells can proliferate via an antigen-independent pathway and may persist in the circulation for a long time before converting to a memory phenotype making this phenotype a less accurate measure of recent thymic emigrants. In SCID patients, very small numbers of mature CD45RO+ T cells predominate for the first 100 days, while thymus-derived naïve CD45RA+ T cells carrying TRECs become the predominant T cells in the circulation from 6 to 12 months after BMT. Paralleling the decrease in TRECs, the number of CD45RA+ cells declines thereafter, although CD45RA+ T cells continue to predominate over CD45RO+ T cells until 10–12 years after transplant .
The T cell receptor repertoire (TCR spectratype) is another measure of the function of the lymphoid compartment beyond simply measures of lymphocyte cell numbers. The TCR heterodimer, formed from rearranged TCR α and β chains, determines the antigen specificity of the T cells according to the variation in the complementarity-determining region 3 (CDR3). Each T cell has a unique TCR with a characteristic CDR3 region. The analysis of CDR3 in a T cell population provides a measure of the diversity of the T cell repertoire and can be performed using a PCR-based technique, or by flow cytometry. T lymphocytes circulating in the blood of normal subjects generally express a widely diverse polyclonal TCR repertoire, which approaches a normal (Gaussian) distribution of the size of CDR3 fragments amplified by primers flanking the variable regions. A skewed TCR spectratype indicates the dominance of a single or a limited number of TCR rearrangements in the T cells. In some long-term surviving XSCID patients with early high TREC levels and a polyclonal T cell repertoire, a skewed repertoire may be evident by 10 years post transplant correlating with the decline in TREC levels and CD45RA+ cells. The oligoclonality was noted primarily in the CD3+CD8+CD45RO+ T cell subset, a phenomenon normally coincident with aging in the normal population . It is unclear whether this rapid decline in thymic function is due to the inability of a SCID patient’s hypoplastic thymus to sustain long-term thymopoiesis or whether donor engraftment with long-lasting progenitor cells was inadequate.
A recent report by Cavazzana et al. observed that patients with SCID who have undergone full myeloablation had donor myeloid cells and persistent thymopoiesis, as indicated by the presence of naïve T cells carrying TRECs [49, 50]. This observation is attributed to the removal of competing immature host thymocytes, thus creating intrathymic stromal space or niches for the incoming donor progenitor cells capable of self-renewal and long-term T cell production. The γc deficiency in XSCID, compared to other SCIDs, seems to leave empty a larger number of stromal niches for donor early progenitor cells to occupy upon transplant. The concept of competition for stromal niches among early T cell progenitors (CD4-CD8-precursor cells) and the utility of irradiation to create niches is supported by the Rag (1/2) knockout mouse model . In contrast to an earlier report by Myers et al. [11.], this study did not find age to be a factor for favorable outcome, whereas the dose of hematopoietic stem cells was found to correlate with improved immune reconstitution . Further underscoring the difficulty in generalizing from studies of different patient groups, 20 of 31 patients (65%) in Cavazzana’s report examined 10–27 years after transplant had positive TRECs, persistent naïve T cells, and overall improved T cell reconstitution, in contrast to the loss of markers of T cell immunity (TRECs, TCR Vbeta diversity) by 14 years as described by Patel et al. [44, 48]. The lack of myeloconditioning in the latter study may well contribute to this disparity and provides further evidence for the benefits of myeloconditioning prior to haploidentical transplants as well as the need for much larger studies to confirm this observation.
Immune deficient patients often have increased susceptibility to gastrointestinal symptoms such as chronic diarrhea, protein loss, and at times abdominal pain or distension. When these symptoms are severe, bloody diarrhea and systemic manifestations in association with GI ulcerations can be indistinguishable clinically from the Crohn’s-like colitis or inflammatory bowel disease reported in other genetic immune deficiencies such as common variable immunodeficiency (CVID), chronic granulomatous disease (CGD), or Hermansky-Pudlak disease [52–54]. Potential etiologies for GI pathology and symptoms in immunodeficient patients include; loss of mucosal barrier integrity, chronic antigen challenge, dysregulation of immunity, underlying defects of the GI mucosa, bacterial overgrowth, and in the case of post-BMT patients, GvHD. We are not aware of published reports of GI problems in these long-surviving partially corrected XSCID patients following a T cell depleted, haploidentical BMT. However, a significant number of the post-haploidentical transplant XSCID patients in their second decade of life referred to us at NIH have chronic diarrhea with varying degrees of protein loss, malabsorption with accompanying iron-deficiency and associated anemia, and fat-soluble vitamin deficiency. Although they are more susceptible to Clostridium difficile infections, other pathogens are not typically isolated from these patients with chronic diarrhea. Malabsorption leading to hypoproteinemia, iron and fat-soluble vitamin deficiency, and difficulty in maintaining serum immunoglobulin levels complicate their medical care. In all 3 patients that underwent endoscopy and GI biopsies at the NIH, there was widespread blunting and occasional absence of villi, with some lymphocytic infiltrates and cryptitis. GvHD or infective etiology was unlikely due to the absence of significant apoptosis, mucosal ulcerations, or positive stains for organisms. Furthermore, antimicrobial therapy has not been beneficial with the exception of eradicating coexisting small bowel overgrowth. Our approach to the management of the GI problems of these patients includes frequent screening for and treatment of intercurrent GI infections or small bowel overgrowth. Malabsorption and nutritional deficiencies such as iron-deficiency anemia is common, and for those who fail to improve on oral iron supplement, we provide regular intravenous iron infusions. In addition to multi-vitamin supplementation, we have found subcutaneous vitamin K injections necessary and sufficient to correct prolonged activated partial thromboplastin times. With severe protein loss, an increase in gamma globulin supplementation to a minimal protective level has to be balanced with the increased cost and morbidity of regular immunoglobulin infusions, either intravenously or subcutaneously. Lastly, a low dose steroid regimen needs to be considered for control of the GI symptoms. In some patients, corticosteroids (oral prednisone or entocort) may help control symptoms and protein loss; however, the complications of long-term steroid use (including osteoporosis, infection risk, and adrenal suppression) require that they be used judiciously.
As mentioned above, HSCT has been a life saving treatment for XSCID and has allowed many survivors to enter adolescence and even adulthood. In many partially corrected patients, profound short stature (below the 5th percentile for age) is a major medical problem, even in those with reasonable nutritional status, and can be an important psychologic concern as well. While chronic illness may contribute to the poor growth in XSCID patients, the severity of the short stature appears to be out of proportion to the clinical history of infection or nutritional status especially when compared to patients with other immune deficiencies that result in frequent infections (e.g., some other forms of SCID, CGD). This raises the possibility that the growth defect observed in these children may be another manifestation of the underlying γc mutation.
Partially corrected XSCID patients that attend school regularly and are socially active are particularly susceptible to communally acquired viral infections typical in these settings. Influenza viruses cause serious (sometimes fatal) disease in immunocompromised patients . Prolonged infection and viral carriage, with the risk of dissemination, and risk of spread in the immune compromised population is well recognized [60, 61]. Persistent B cell dysfunction in XSCID patients following T cell depleted, haploidentical BMT results in an inability to mount specific antibody production in response to antigen challenge. Killed influenza vaccines are less effective and live attenuated vaccines are contraindicated due to the risk of developing severe disease in these patients. IVIG offers some protection if the donor pool has antibodies of relevant specificity to circulating strains. While some promising antiviral therapies are being developed, for example neuraminidase inhibitors, they are most effective at the onset of illness. Due to the underlying immunodeficiency in these patients, classical flu symptoms are seldom detected to allow prompt diagnosis and treatment thus limiting their use unless the patient is aggressively monitored and treated.
Titration of nasal washings was relatively laborious but it provided a helpful estimate of therapeutic efficacy and informed our choice to augment the treatment with additional NI. Aggressive management of infections in these patients is of particular importance due to several factors including prolonged viral shedding , and increased risk of evolution of drug resistance , which may further increase the risk of nosocomial spread . Side effects from oseltamivir may also limit its long-term use or use at recommended dosages. In this patient, addition of a second NI may have aided the elimination of the virus as the titers rapidly decreased following the addition of zanamivir (Fig. 3). This case is presented to highlight the importance of suspecting and screening for influenza infections, and the importance of aggressive monitoring of viral load in immuno-compromised patients to document viral eradication. A year later, influenza A was isolated when he again presented with fever and rhinorrhea, which resolved following prompt treatment with oseltamivir. Three months later, at first onset of rhinorrhea alone, a prompt nasal washing was performed and parainfluenza3 isolated. He improved following 2 weeks of inhaled ribavirin. Of note, with aggressive and prompt treatment, this patient was able to eradicate the virus on each occasion. This patient has since proceeded to a successful, matched, unrelated cord blood transplant.
Severe cutaneous papillomavirus disease is a problem in some partially corrected XSCID patients following HSCT  that may require more than conventional therapy. This problem has been reported at multiple centers and in over 20% (9/41) of patients at least 10 years after HSCT for SCID due to γc or Jak3 deficiency. Furthermore, XSCID dogs [25, 64] are sensitive to papilloma virus which frequently progresses to squamous cell carcinoma . Of 24 XSCID dogs followed for at least 1 year post-BMT, 71% developed chronic canine papillomavirus infection. Six of the transplanted dogs that developed cutaneous papillomas were maintained more than three and a half years post-BMT for use as breeders. Four of these six dogs developed invasive squamous cell carcinoma (SCC), with three of those progressing to metastatic SCC, an extremely rare consequence of SCC in the dog. In many HSCT-treated XSCID patients, a persistent abnormality in NK cells or persistent γc defects in keratinocytes may explain the susceptibility to HPV [63, 66]. Thus, a persistent quantitative deficiency of NK cells after hematopoietic stem cell transplantation could prevent efficient anti-HPV immunity. Treatment modalities that were used include laser treatment, acitretin, and topical imiquimod that resulted in complete remission. All other patients had received more than one treatment for HPV disease with limited efficacy. Two patients received a boost of donor hematopoietic stem cells that led to sustained remission in only one patient.
Chronic lung disease
Chronic lung disease with bronchiectasis is not uncommon in these partially corrected XSCID patients. Major contributors to chronic lung disease include; severe pulmonary infections prior to transplant, GvHD, recurrent bronchopulmonary infections despite IVIG, and persistent B cell dysfunction. These patients may present with bronchiectasis on CT scan, frequent positive sputum or BAL cultures, and markedly impaired pulmonary function on spirometry. In some, their pulmonary status may be the limiting factor to their quality of life, or even the determining factor for survival. Due to this severity of pulmonary involvement, lung transplants have been considered for some of our patients, although there is little experience with such transplants in patients with coexisting immune deficiency. Unfortunately, poor pulmonary status is a bad prognostic indicator for BMT. Our limited management strategies for the pulmonary problems in these patients include aggressive treatment of pulmonary symptoms as guided by identification of pathogens in sputum or BAL. In some patients, a rotating regimen of oral and/or nebulized antibiotics have kept pulmonary status relatively stable for the short term. When bronchiolitis obliterans is present, there is little documented effective therapy, with the exception of high dose steroid . For the long term, more definitive treatment options need to be considered as discussed below.
While haploidentical bone marrow transplants can save lives in patients with XSCID, follow-up of long-term survivors reveals the emergence of several complicated problems in a subset of patients for which there are limited treatment solutions other than pursuing functional immune reconstitution. Hematopoietic stem cell transplant in this subpopulation of previously transplanted patients presents high risks due to their impaired health status typically including existing organ damage. Furthermore, the presence of donor T cells requires aggressive myeloconditioning in order to minimize graft versus graft rejection. In this setting, gene therapy is a reasonable consideration to avoid intense chemotherapy.
Success in infants
Gene therapy can achieve a remarkable immune correction in XSCID infants . Nine of ten patients in the initial clinical trial were effectively cured of their immuno-deficiencies after reconstitution with genetically corrected lymphocytes derived from transduced, autologous bone marrow—repopulating cells . A second clinical trial also reported a positive outcome with restoration of functional immunity in four young children with XSCID. Thus, substantial immunological reconstitution has been observed in 19 infants in two similar clinical trials using a long-terminal repeat (LTR)-regulated γ-retroviral vector carrying functional γc. Overall, in gene therapy for genetic immunodeficiencies, more than 80% of all patients with these disorders treated by gene therapy during the past 5 years have had a lasting benefit. There are excellent reviews on gene therapy, and the following section gives a brief overview of some of the critical elements and an update on recent developments [70–74].
Failure in older children
In contrast to the experience with infants, older patients have generally but not uniformly proven to be quite resistant to achieving gene marking using current ex vivo, retroviral vector mediated strategies. Our group  and another in London  have treated five older XSCID patients with ex vivo gene therapy [18, 50, 75]. While similar vectors have successfully treated animals  and infants [69, 75], only one older patient of the five has achieved stable, high levels of γc gene marking [18, 50]. At 3 years after gene therapy, his T cell number has not significantly increased, despite over 90% of his T cells remaining marked. This highlights the unique problems specific to the older XSCID patients. Abnormal thymopoiesis in these older children is likely a contributing factor. Of note, success at gene therapy for patients with ADA SCID similarly encountered poor levels of marking until cytoreductive conditioning agents were added to remove resident T cells, thus creating ‘space’ in the hematologic compartment for the incoming gene-modified autologous cells. Similarly, in older patients, the addition of cytoreductive agents may be necessary to enhance engraftment of the autologous, gene corrected cells. Such modifications to add cytoreductive measures to current protocols have been made but outcomes are pending.
While gene therapy is a potentially powerful treatment for XSCID and other disorders, there are intrinsic risks. In 4 out of 9 patients from the first study, insertion of the gene therapy vector resulted in upregulation of host cellular oncogenes and consequent leukemogenesis within 3–6 years of gene therapy [77–79]. In a second XSCID gene therapy trial, one out of 10 patients developed leukemia after 24 months . Furthermore, oncoretroviral gene therapy for CGD, a myeloid immune deficiency, also caused outgrowth of clones with the majority of myeloid cells containing insertions into Evi1-MDS1 following an initial polyclonal pattern . These events suggest that conventional murine γ-retroviral vectors (based on MLVs) present a high risk of insertional mutagenesis and proto-oncogene activation as a result of their powerful enhancer sequences in the intact long-term repeat (LTR) regions.
Potential etiologies for lymphoproliferative disorder
Remarkably, 4 out of the 5 patients in the 2 XSCID gene therapy trials who developed lymphoproliferative disease have an activating insertion in the LMO2 locus [82, 79]. This indicates that either this is a preferred site for integration of γ-retroviral vectors or that integration at this site imparts a powerful preleukemic predisposition (or both).
Oncoretroviral vectors have demonstrated a propensity to integrate near promoter regions with 20–25% of all integration events occurring within 10 kb of transcriptional start sites, particularly in actively transcribed genes [83, 84]. In cell culture, these integration events cause immortalization or permit proliferation independent of exogenous growth factors . Uncontrolled growth, or tumors, found in animals following gene therapy may result from destructive insertions into tumor suppressor genes, pro-apoptotic genes, (BCL2A1), regulator of tumorigenesis (Evi1, MDS1, or PRDM16) or by activating insertions in proto-oncogenes, positive growth regulators, or anti-apoptotic genes [86, 87]. Certainly in the XSCID patients that developed leukemia following gene therapy, retroviral insertion into LMO2 resulted in overexpression of the allele into which the insertion occurred . LMO2 is a known proto-oncogene associated with T cell leukemia . The enhancer elements in the LTR affects the disease spectrum in vivo, and the vector in both XSCID trials contains the Moloney murine leukemia virus (MoMLV) known to cause thymic lymphomas in newborn mice of the NFS strain . Although the expression of γc and its downstream signaling molecule is not increased following gene therapy in XSCID, a synergistic effect between γc and LMO2 is not excluded [82, 90]. Of interest, LMO2 over-expression in transgenics and retrovirally mediated γc in LMO2-transgenic mice do not appear to result in development of lymphomas , suggesting perhaps the requirement for a ‘second hit.’
Proposed improvements to safety, efficacy, and testing of gene therapy vectors
Various improvements in safety and efficacy have been reported recently although the occurrence of adverse events may require significantly more patient years to become evident [92–94]. The choice of the viral source of the vector backbone is of great importance. Overall, lentiviral vectors possess several attractive features, notably their patterns of integration  and their ability to transduce quiescent cells which decreases the need for extensive ex vivo culture and stimulation  as well as the presence of both a self-inactivating (SIN) mechanism [96, 97] and a chromatin insulator . While MLV’s preferred integration is near the start of transcriptional units (either 5′ or 3′), lentiviral vectors favor intragenic sites with a distribution along the entire length of the gene . The propensity for distribution throughout the entire gene is also true of foamy virus vectors . The lack of bias toward integration near promoters in lentiviral vectors should favor a reduction in the disruptive effects that integrants may have on the promoters, as demonstrated by the ‘promoter trapping’ studies .
While the 3rd generation lentiviral system incorporated the SIN feature many years ago [96, 101], the recent addition of a SIN feature to an oncoretroviral vector demonstrated multilineage lymphoid reconstitution in a murine XSCID model at a similar level to that achieved by a conventional LTR-regulated vector used in previous clinical trials . The SIN system incorporates a deletion in the U3 region which, when transcribed, copies over to the 5′-end upon proviral integration, resulting in inactivation of the viral promoter/enhancer elements following reverse transcription. This results in a system in which the LTRs of the integrated provirus are devoid of enhancer and promoter sequences. In the conventional oncoretroviral system, both the 5′ and the 3′ LTRs can promote read-through transcription of downstream sequences, with potential interaction between the enhancer elements within the LTR and neighboring promoters. Thus, SIN vectors rely on a single internal promoter to drive transcription of the transgene [102, 96, 97], which theoretically should attenuate the interaction with neighboring cellular genes thus reducing toxicity. While the transforming capacity of SIN vectors is significantly reduced when compared with corresponding LTR vectors, strong internal enhancers within SIN vectors remain capable of transforming cells by insertional mutagenesis and may trigger clonal imbalance [85, 103].
An additional safety feature is the addition of insulator elements which act as a barrier between heterochromatin and transcriptionally active chromatin [94, 98] and may thus reduce interactions between enhancers and neighboring genes. We and others have found that use of the chicken HS4 β-globin insulator results in more homogenous and consistent stable protein expression in transduced hematopoietic stem cells [104, 105]. Retroviral enhancer–promoters also have a significantly greater potential to activate neighboring promoters than cellular promoters derived from human genes such as elongation factor-1 (EF1) and phosphoglycerate kinase (PGK) , as implicated by the lack of activation of the crucial proto-oncogene Evi1 activation despite integration of multiple copies per cell, and were not detectable when using SIN-EF1 vectors. The decrease in transforming activity was significantly different regardless of the presence of a well-characterized genetic insulator core element .
In considering issues related to efficacy, oncoretroviral vectors require cells to be actively dividing for integration. Thus, cytokine pre-stimulation of target hematopoietic stem cells is used to drive proliferation. In contrast, lentiviral vectors are able to transduce quiescent non-mitotic cells, which may theoretically reduce ex vivo manipulations of HSCs known to alter cell homing and engraftment properties . A recent study using a lentiviral vector demonstrated an improvement in the efficacy of stem cell transduction in a non-human primate model, and it is probable that similar results could be obtained in humans. A lentiviral gene therapy vector has been employed to treat adrenoleukodystrophy and preliminary data have shown encouraging levels of marking in all lineages [108, 109].
Tools for evaluating future clinical vectors for gene therapy
Insertional mutagenesis was primarily a theoretical concern despite preclinical studies with animals. Much progress has been made to address the issues raised by the recent clinical trials in the design of vectors, the methods of transduction and ex vivo culture conditions. Central to this development is the excellent progress made in the design of both in vitro and ex vivo models for testing insertion-related adverse events, thus allowing improved pre-clinical assessment of future clinical vectors. In vitro ‘immortalization’ systems, or assays based on in vitro expansion of primary murine hematopoietic cells evaluate the efficiency with which the integration of a provirus can lead to growth dysregulation [85, 87]. Mouse models with critical elements removed to lower the threshold for development of tumors , or to create an environment recapitulating that of insertional mutagenesis in XSCID, have also been created to provide in vivo assays for evaluating the safety of gene therapy vectors.
In the partially corrected, older XSCID patients, there are a number of additional barriers to successful gene therapy and cure. Engraftment of stem cells is less efficient with increased age (as seen in HSCT) in combination with poorly developed thymus. Moreover, the presence of immune competent donor T cells from previous transplants may increase risk of rejection and may require greater cell doses per weight basis. Because of the successes and failures of the current state of the art gene therapy approaches, the next generations of gene therapy will have to possess improved safety features validated in well-designed models for genotoxicity. Myeloconditioning agents may help to remove grafts and clear niches in the myeloid, lymphoid, and perhaps thymic compartments. Of course the risk/benefit analysis of gene therapy approaches will have to constantly be balanced against developments in other treatment modalities. MUD transplant, a procedure that undoubtedly would require intense preparative chemotherapy and GvHD prophylaxis, is also improving and remains worthy of much further study. Finally, as the various therapies described above help to improve the immune status of these patients, the astute clinician and patient must together be wary of new manifestations of the underlying clinical problems experienced by these patients.
This project was supported by the Intramural Research Program of the NIH, National Institute of Allergy and Infectious Diseases. We thank Dr Kol Zarember for his critical review of the manuscript and apologize to the authors of many excellent publications that we were not able to present adequately in this review.