| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Crucell B.V., Leiden, The Netherlands
Correspondence: D. W. van Bekkum, Crucell B.V., P.O. Box 2048 2301 CA, Leiden, The Netherlands
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRELEVANCE OF ANIMAL MODELS...ADJUVANT ARTHRITIS IN BUFFALO...EXPERIMENTAL AUTOIMMUNE...AID AS DISEASES OF...TREATMENT OF FULLY DEVELOPED...AUTOLOGOUS BMTEXPERIMENTAL DATA OF...RECOMMENDATIONSREFERENCES |
|---|
Key Words: transplants • bone marrow • arthritis • encephalomyelitis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRELEVANCE OF ANIMAL MODELS...ADJUVANT ARTHRITIS IN BUFFALO...EXPERIMENTAL AUTOIMMUNE...AID AS DISEASES OF...TREATMENT OF FULLY DEVELOPED...AUTOLOGOUS BMTEXPERIMENTAL DATA OF...RECOMMENDATIONSREFERENCES |
|---|
The clinical literature contains sporadic case reports of the transfer of AID and allergic disorders with donor BM to patients treated for leukemia or aplastic anemia. These include thrombocytopenic purpura, thyroiditis, diabetes type I, celiac disease, and myasthenia gravis [2 ]. After the discovery that full-blown AID of experimental animals can be cured by allogeneic BM transplantation (BMT), the records of patients who were long-term survivors of treatment with allogeneic BMT for leukemia or aplastic anemia were searched for those with coexisting AID at the time of transplantation. The search revealed 21 patients suffering from rheumatoid arthritis (RA), psoriasis, Crohn’s disease, ulcerative colitis, SLE, or insulin-dependent diabetes type 1 (IDD1), all of whom experienced a complete remission of their autoimmune disorder. The transplant-associated risks of allogeneic BMT, however, precluded its use in AID. It was only after our unexpected finding that autologous BMT could also cure experimental AID that transplantation of BM or peripheral blood stem cells (PBSC) were introduced into the clinic as a treatment option for refractory AID. Thus far, the results of these clinical studies, which comprise several hundreds of patients, have established a highly predictive value of the disease models that were used.
|
RELEVANCE OF ANIMAL MODELS FOR HUMAN AID |
|---|
|
TOPABSTRACTINTRODUCTIONRELEVANCE OF ANIMAL MODELS...ADJUVANT ARTHRITIS IN BUFFALO...EXPERIMENTAL AUTOIMMUNE...AID AS DISEASES OF...TREATMENT OF FULLY DEVELOPED...AUTOLOGOUS BMTEXPERIMENTAL DATA OF...RECOMMENDATIONSREFERENCES |
|---|
The animal models of AID are of two distinct categories: the spontaneous (or hereditary) and the induced forms. In the first category, the disease develops spontaneously in a large proportion or in all of the individuals of a so-called autoimmune strain of mice or rats. Well-known examples are the lupus-like syndromes in several inbred mouse strains and diabetes in NOD mice and BB rats.
The induced AID require immunization with certain antigens to develop and do so only in selected inbred strains—the susceptible ones or responders—and not in others—the resistant or nonresponder strains. The best described models are adjuvant arthritis (AA) in rats and experimental allergic encephalitis (EAE), which can be induced in many species. The latter is considered the most appropriate animal model for MS. Susceptibility and resistance to inducible AID have been shown to be genetically determined by cross-breeding experiments.
A continuous subject of debate is the question of what type of disease model is expected to best predict efficacy of new therapeutic modalities in human AID. The induced models appear to be the favored candidates, as they share a dual ethiology with their human counterparts. However, some of the spontaneous AID of animals have also been found to be strongly influenced if not dependent on environmental factors, such as microbes, food, and hormones. For instance, HLA-B27 transgenic rats do not develop the characteristic colitis and arthritis when reared germ free [5 ]. In germ-free NZB mice, the incidence and severity of renal lesions were much lower than in conventional controls [6 ], but in MRL-lrp mice similar experiments did not provide any evidence for a role of infectious agents in the development of their lupus-like disease [7 ]. Exposure to a common rat virus, the Kilman rat virus, initiates autoimmune IDD1 in some but not all rat strains that are normally not susceptible to spontaneous diabetes [8 , 9 ]. On the other hand, elimination of environmental viruses by Caesarian derivation of diabetes-prone BB rats increased the incidence and rate of development of diabetes [10 ]. Other environmental factors that have been identified in diabetes-prone rodents are the diet and hormones. A semi-purified diet prevents diabetes in BB rats [11 ], and a raised environmental temperature, which decreases food intake, reduced the incidence of diabetes in NOD mice [12 ]. Exposure to environmental stressors enhanced the onset and increased the incidence of diabetes in BB rats [13 ], and sex hormones influence the occurrence of insulinitis and diabetes, sialitis, and dacryoadenitis in NOD mice [14 ].
Interestingly, in several instances, the microflora or hormones was also found to modify the susceptibility to induction of AID. F344 rats are resistant to induction of adjuvant arthritis under conventional conditions but are highly responsive in the germ-free state [15 ]. This was not the case for collagen-induced arthritis (CIA), to which F344 rats are equally resistant under both conditions. However, CIA was markedly enhanced in germ-free (susceptible) DA rats as compared with conventional ones [16 ]. In this context, it should be reminded that graft versus host disease (GvHD), which has many pathological features in common with SLE and autoimmune colitis, is strongly influenced by the composition of the microflora in experimental animals [17 ] and human patients [18 ].
The disease pattern of EAE is profoundly influenced by the plasma level of corticosteroids. Nonresponsive PVG rats develop severe and fatal EAE when they are adrenalectomized prior to induction. Adrenalectomy of the responder Lewis rat alters its usual pattern of acute remitting, nonrelapsing EAE into nonremitting fatal disease [19 ]. The original reaction pattern is restored by replacement therapy with corticosteroids. Rats of the low-responder strain WAG have an incidence of 10% following immunization, which increases to 50% in adrenalectomized animals [20 ].
Another factor identified as influencing EAE development is the presence of cyclophosphamide (CY)-sensitive suppressor cells. Abrogation of the resistance of acridine orange rats [21 ] and of BALB/c mice [22 ] to encephalitogenic immunizations was obtained by pretreatment with low dose (20 mg/kg) CY. The incidence of subclinical EAE as defined by histological lesions in the central nervous system (CNS) of WAG rats increased from 37% to 86% by pretreatment with a similarly low dose of CY, but the incidence of clinical EAE remained unaltered at 10% [21 ].
In view of these data, the differences between the spontaneous and the inducible animal models may well be gradual rather than fundamental. In both cases, there is a genetic disposition that allows activation of antiself immunity; in the case of the spontaneous diseases, activation is by relatively weak, ubiquitous antigens; in the case of the inducible AID, by strong, specific antigens. Both mechanisms accommodate a role for infectious microorganisms acting as the initiating stimulus by providing antigens that resemble tissue targets (so-called mimicry) or as regulators of the immune reactivity, similar to adjuvants, hormones, and dietary factors. All these determinants, genetic as well as environmental, have also been implicated in the pathogenesis of human AID. The etiology of human AID is undisputedly multifactorial. In almost all autoimmune conditions, there is a familial tendency. Many AID are induced by drugs: more than 70 different drugs have been reported to induce SLE. Furthermore, many xenobiotics, e.g., food supplements, heavy metals, and environmental toxins, have been linked to the development of SLE-like illnesses. Although in many patients with AID, the causative agent remains unknown, the low concordance in identical twins seems to argue in favor of the induced disease models as being more realistic tools for translational research.
|
ADJUVANT ARTHRITIS IN BUFFALO (BUF) RATS |
|---|
|
TOPABSTRACTINTRODUCTIONRELEVANCE OF ANIMAL MODELS...ADJUVANT ARTHRITIS IN BUFFALO...EXPERIMENTAL AUTOIMMUNE...AID AS DISEASES OF...TREATMENT OF FULLY DEVELOPED...AUTOLOGOUS BMTEXPERIMENTAL DATA OF...RECOMMENDATIONSREFERENCES |
|---|
|
EXPERIMENTAL AUTOIMMUNE ENCEPHALOMYELITIS IN RODENTS |
|---|
|
TOPABSTRACTINTRODUCTIONRELEVANCE OF ANIMAL MODELS...ADJUVANT ARTHRITIS IN BUFFALO...EXPERIMENTAL AUTOIMMUNE...AID AS DISEASES OF...TREATMENT OF FULLY DEVELOPED...AUTOLOGOUS BMTEXPERIMENTAL DATA OF...RECOMMENDATIONSREFERENCES |
|---|
Relapsing EAE develops after a similar immunization in SJL/J mice, in Biozzi mice, in BUF rats, and in Lewis rats, but in the latter only if pretreated with low-dose cyclosporin A. Relapses are characterized by one or more new episodes of paralysis and paresis after the animals have recovered from the first attack. Most of the animals recover after each subsequent attack; some 10% of the BUF rats succumb during that period. The spontaneous relapses are accompanied by more widespread inflammation in the CNS and more pronounced demyelination. After 40–50 days, relapses no longer develop spontaneously, but can be reinduced by another immunization. Such induced relapses are likely due to activation of memory T cells, as the latent period is 2–6 days less than in the case of primary immunization. It remains a matter of speculation how to interpret spontaneous and induced relapses as models for relapses that occur in MS patients. In all cases, the relapses are associated with flare-up of the inflammatory processes in the CNS at "old" sites or at new locations. As it is unknown what causes relapse in MS, notably if it represents a nonspecific activation or a re-exposure to the sensitizing antigen, it cannot be decided whether induced relapses in the animals have any counterpart in the clinic. Nevertheless, I and others have recorded the incidence of induced relapses after BMT, reasoning that it will provide information on the presence of residual memory cells.
In the Biozzi mouse, immunization with spinal cord tissue induces a chronic form of EAE, which eventually leads to more extensive demyelination than is observed in the remitting relapsing EAE of rats. EAE in Biozzi mice is therefore a better model for studying the effects of therapeutic interventions on chronic, persistent demyelating disease, which is mainly due to the scars from previous inflammations.
|
AID AS DISEASES OF THE HEMATOPOIETIC STEM CELLS |
|---|
|
TOPABSTRACTINTRODUCTIONRELEVANCE OF ANIMAL MODELS...ADJUVANT ARTHRITIS IN BUFFALO...EXPERIMENTAL AUTOIMMUNE...AID AS DISEASES OF...TREATMENT OF FULLY DEVELOPED...AUTOLOGOUS BMTEXPERIMENTAL DATA OF...RECOMMENDATIONSREFERENCES |
|---|
An alternative hypothesis is based on the finding that the potential for developing spontaneous AID can be transferred by BMT to lethally irradiated animals from a nonautoimmune strain, and reversely, autoimmune-prone animals do not develop the disease when grafted at an early age, before the disease is manifest, with BM from a normal strain. This property was discovered in NZB mice, which spontaneously develop a lupus-like syndrome, by Morton and Siegel [1 ] who postulated "that some defect...may exist in the NZB mouse at the level of the hemopoietic stem cell." Similar transfers were demonstrated in several different spontaneous AID strains as listed in Table 1 . It also led to a re-emphasis, most strongly by Ikehara et al. [30 ], of AID being stem cell diseases. However, this conclusion is not fully justified as long as the transfers have not been accomplished with highly purified stem cells. For the time being, it seems appropriate to use the term "stem cell-associated diseases."
|
|
SWR chimeras, which might be a result of the C5 deficiency of the recipients. Surprisingly, the SWR DBA/1 chimeras were susceptible. In this particular case, the inherent potential to react against collagen may be present in the so-called resistant donor strain but masked by its C5 deficiency.
In the EAE group of experiments, those reported by Korngold et al. [47
] are contrasting with the other data. In their paper, they demonstrate that the responses of the chimeras are dictated by the recipient strain, similarly when the CNS antigen used for induction was of BALB/c or B10.S origin. In another publication of the same year, Lublin et al. [48
] confirmed these results with the SJL-derived antigen, but immunization with the B10.S antigen induced a high incidence of clinical EAE in B10.S SJL and SJL B10.S chimeras. A satisfactory explanation has not emerged so far.
|
TREATMENT OF FULLY DEVELOPED EXPERIMENTAL AID WITH ALLOGENEIC BMT |
|---|
|
TOPABSTRACTINTRODUCTIONRELEVANCE OF ANIMAL MODELS...ADJUVANT ARTHRITIS IN BUFFALO...EXPERIMENTAL AUTOIMMUNE...AID AS DISEASES OF...TREATMENT OF FULLY DEVELOPED...AUTOLOGOUS BMTEXPERIMENTAL DATA OF...RECOMMENDATIONSREFERENCES |
|---|
|
|
The chronic disease is characterized by extensive demyelination, axonal loss, and gliosis. The majority of the remitting animals experience one or more relapses. Mice treated with allogeneic BMT when chronically ill at day 108 post-immunization did not respond at all to allogeneic BMT. Apparently, treatment with BMT effectively halts the inflammatory processes but does not lead to repair of scar tissue in the nervous tissue.
Overall, the results obtained with allogeneic BMT are impressive, considering that the animals were quite sick when treated and had almost no relapses. These findings stimulated a search of the clinical records of long-term survivors of allogeneic BMT for patients with coexisting AID at the time of grafting. Twenty-one patients with a follow-up of 7–21 years were discovered. All experienced complete remission of their various AID and only one patient relapsed. These cases have been reviewed by Marmont [61 ] and Nelson et al. [62 ].
Despite such strong evidence of its therapeutic potency, allogeneic BMT has not been intentionally applied to the treatment of severe AID so far, except for refractory, idiopathic, aplastic anemia, where it is a long-established therapy, primarily intended to replace the lost BM. The highly successful treatment of this life-threatening AID with allogeneic BMT is perhaps the most convincing argument for extending this modality to other AID. However, the considerable risk of transplantation-associated mortality has precluded this.
The discovery that autologous BMT is equally effective as allogeneic BMT in inducing complete remissions in rats with AA an EAE cleared the way for clinical application, as autologous BMT is less risky. Among more than 500 patients registered so far as having been treated with autologous stem cells for severe refractory AID, the overall transplant-associated mortality was 9%, with significant variation between diseases [63 ].
|
AUTOLOGOUS BMT |
|---|
|
TOPABSTRACTINTRODUCTIONRELEVANCE OF ANIMAL MODELS...ADJUVANT ARTHRITIS IN BUFFALO...EXPERIMENTAL AUTOIMMUNE...AID AS DISEASES OF...TREATMENT OF FULLY DEVELOPED...AUTOLOGOUS BMTEXPERIMENTAL DATA OF...RECOMMENDATIONSREFERENCES |
|---|
Curiosity made us investigate treatment with autologous BM. Once more, against all expectations, autologous BM was just as efficacious as syngeneic marrow from healthy animals and as allogeneic marrow [64 ]. These results were confirmed with grafts of real autologous BM, which was harvested from the femur of arthritic recipients by a surgical procedure, followed by total body irradiation (TBI) and intravenous return of their own BM cells [64 ].
Subsequent studies with autologous stem cells were always performed with BM harvested from animals with exactly the same stage and severity of the disease as the recipients. The marrow obtained in this way was termed pseudoautologous at that time. By this procedure, unnecessary suffering of the very sick animals from the surgical intervention needed for BM collection is avoided. For each experiment, about 100 rats were immunized, and when the disease was fully developed, each animal was scored using a grading scale for the clinical symptoms. The animals were distributed over the various experimental groups and the donor group of rats, assuring that the average score of all groups was similar. Animals without symptoms (10–20%) were always excluded. As the composition and properties of pseudoautologous and autologous BM are identical, the term autologous is used for both throughout this review.
In contrast to the preclinical experiments with autologous BMT, those with syngeneic transplants have little practical significance, as identical twin donors are rarely available. However, any difference in the results of treatment between autologous and syngeneic cells is of great interest, as it may shed light on the significance of activated T cells and memory cells in the graft. As can be seen from Table 5 , few published data are available on syngeneic and autologous BMT in the treatment of spontaneous AID. An early publication of Morton and Siegel [27 ] contains an experiment with syngeneic BMT in lethally irradiated 6-month-old (NZBxDBA/2) F1 mice. At that time, 45% of the recipients were positive for antinuclear antibodies. This proportion did not decline after the treatment, in contrast to the near complete disappearance after allogeneic BMT. The study did not contain a description of the clinical condition of the mice. The second publication [40 ] concerns the failure to cure HLA-B27 transgenic rats with syngeneic BMT, as contrasted with the successes with allogeneic BM in the same model. Good and Ikehara [70 ] may have investigated syngeneic and autologous BMT in the course of their extensive studies on the treatment of spontaneous, immune disorders, as they stated, "Our preclinical studies do not support autologous or syngeneic BMT for treatment of mice, which may already have developed systemic autoimmune disease." However, details of such experiments have not been published.
|
Karussis et al. [71 ] treated MLR/lpr mice, aged 9–10 weeks, with high-dose CY or 9 gray (Gy) TBI followed by syngeneic BMT. These mice remained disease-free for at least 36 weeks, and untreated controls began to die at week 16. Unfortunately, it is not stated whether the recipients were already sick at the time of transplantation nor were the age and disease status of the BM donors reported.
Loor et al. [72 ] reviewed several reports of complete and lasting remissions in mice with spontaneous lupus-like diseases after treatment with sublethal TBI. Furthermore, long-term treatment with relatively high dose CY (100 mg/kg weekly for 16 weeks) prolonged survival of sick MRL/lpr mice, caused reversal of adenopathy, and prevented the development of arthritis and glomerulonephritis [73 ]. In the latter study, the follow-up was only 6 weeks after the last administration of the drug, which is not long enough. Nevertheless, sublethal TBI and high dose CY cause massive destruction of the lymphatic and hematopoietic cells followed by endogenous repopulation from primitive precursors. That process closely resembles the repopulation following lethal TBI and autologous BMT.
Much more exhaustive investigations were performed with autologous and syngeneic BMT in rodents suffering from induced AID. Excellent therapeutic effects were seen in all studies except for the one with CIA mice. Failure of these mice to enter complete remission was also observed after allogeneic BMT as discussed before.
The studies performed by Karussis et al. [69 , 74 ] on EAE induced in SJL mice are of particular interest, as they also treated the mice before the appearance of clinical disease with high dose CY (300 mg/kg) and syngeneic BM. The induction schedule consisted of two immunizations, 1 week apart. The latent interval between the first immunization and appearance of symptoms was between 14 and 18 days. When treated at day 6 after the first immunization, the development of the disease was completely prevented, and the mice became resistant to rechallenge with the encephalitogenic agent. However, treatment at day 9 after the first immunization was not preventive; it only delayed the onset of paralysis by about 1 week. The authors’ interpretation is that shortly after the second inoculation, the population of autoreactive and memory lymphocytes is at a maximum and cannot be sufficiently reduced by the conditioning with CY. Yet this explanation is difficult to reconcile with the finding that the same regimen was highly effective when applied 3 days after the appearance of clinical disease.
In addition to the results with treatment of spontaneous and induced AID, as collected in Table 5 , we should mention the treatment of adoptively transferred EAE with syngeneic BM by Burt et al. [75 ]. This is a relapsing form of EAE that can be evoked in SJL/J mice with lymph node cells from sensitized, syngeneic donors. The cells are stimulated in vitro with the disease-initiating peptide, proteolipid protein 139–151 prior to transfer. These mice were treated at the peak of the acute phase of the disease at 14 days after transfer or at day 74 during the chronic phase with a lethal dose of TBI or with TBI and cyclophospamide followed by rescue with syngeneic BM. Treatment in the early phase caused a significant clinical and histological improvement, but there was no effect of treatment of the chronically ill animals. This is reminiscent of the lack of responses of chronically ill Biozzi mice to treatment with allogeneic BM and supports the rule that BMT cannot repair scar lesions.
|
EXPERIMENTAL DATA OF TRANSLATIONAL SIGNIFICANCE |
|---|
|
TOPABSTRACTINTRODUCTIONRELEVANCE OF ANIMAL MODELS...ADJUVANT ARTHRITIS IN BUFFALO...EXPERIMENTAL AUTOIMMUNE...AID AS DISEASES OF...TREATMENT OF FULLY DEVELOPED...AUTOLOGOUS BMTEXPERIMENTAL DATA OF...RECOMMENDATIONSREFERENCES |
|---|
Detailed results of autologous BMT in AA and EAE
After high dose TBI (9–10 Gy), autologous BMT causes remissions of both diseases in all animals. In AA, 70% are complete responders, and 30% are partial responders. Spontaneous relapses or exacerbations are extremely rare; also relapses are hardly ever inducible (Table 6
). In this disease model, the outcome seems to be dominated by the intensity of the conditioning. Even the addition of large numbers of autologous spleen cells or lymphocytes from the lymph nodes or from the peripheral blood did not adversely influence the responses nor did it evoke relapses [66
]. So far, this cannot be explained but it seems to be in line with our failure to passively transfer AA in BUF rats with lymphoid cells.
|
This illustrates the dilemma of the translator: is T cell depletion to be recommended for clinical transplants or only for the treatment of patients with MS? As will be explained later, the decision in this case can be made on more pragmatic grounds.
Interestingly, the spontaneous relapse incidence in EAE was only 5% following allogeneic BMT as compared with 30% after autologous and syngeneic BMT [50 , 51 ]. The difference is ascribed to a graft-versus-autoimmunity reaction by which residual, autoreactive cells of the recipient are eliminated, a term introduced by A. M. Marmont in reference to the well-known graft-versus-leukemia effect of allogeneic BM grafts. The analogy with the treatment of leukemia may be extended further by proposing that the treatment of severe AID should also aim at eradication of as many autoreactive cells as possible by conditioning and in case of autologous BM, by ex vivo purging of the graft prior to reinfusion.
The conditioning regimen
In both models, the best results have been obtained with the strongest lymphomyeloablative regimens, e.g., the highest tolerated dose of 9–10 Gy of TBI [66
, 68
]. Partial body irradiation of the affected tissues only (the CNS in the case of EAE or the legs and tail in the case of AA) or shielding of those parts while irradiating the rest of the body resulted, at best, in a limited, temporary remission [58
, 67
]. Fractionated irradiation was investigated in the AA model [66
] and proved to be as effective as single dose TBI, provided the total dose was properly adjusted upward. In AA and EAE, CY alone or busulfan (BU) alone at highest tolerated doses was less effective than high dose TBI, and the combined regimen of CY and BU was equally effective. The combination of a lower dose of TBI (4 Gy or 7 Gy) with a lower dose of CY (2x60 mg/kg) was also as effective as the highest dose of TBI. In experimental autoimmune myasthenia gravis (EAMG), conditioning with CY alone was inadequate, and a moderate dose of TBI had to be added to get complete responses [65
]. Notable features of high dose CY as the sole conditioning agent in AA were not only the lower rate of complete remissions, but also the substantial incidence (36%) of spontaneous exacerbations. In contrast, among 155 AA animals treated with high-dose TBI or the combination regimens, only one relapse occurred.
So far, conditioning of more than 70 patients with severe RA has consisted of high dose CY, either as the sole agent or combined with anti-thymocyte globulin (ATG). Although the majority of the patients went into complete or partial remission, approximately two-thirds relapsed, usually within 1 year [63 ]. It seems fair to conclude that the equivocal results with experimental arthritis have predicted this outcome. The main reason for the rheumatologists not to adopt the combination of CY and moderate dose TBI, which is clearly superior in AA and EAE, has been fear of higher toxicity and mortality. By drawing once more on the analogy with the treatment of leukemia, the relapse rate can only be reduced by intensifying the conditioning. The cost of the latter is more toxicity, unless target cell specificity can be improved upon.
The most likely candidate target cells in overt AID are activated T lymphocytes and memory T lymphocytes. The characteristics of these cell types are still poorly defined.
Yet, there are some indications for differences in target specificity between radiation and CY. In treating rats suffering from EAMG, Pestronk et al. [65 ] showed CY to be less effective than TBI in eliminating immunological memory. In this case, the cells involved were most likely B memory cells, but T memory cells (in this case against M. tuberculosis) were also reported to be CY-resistant [76 ].
In addition, there is clinical evidence that CY as a single agent is inadequate for ablation of memory T cells. In nonsensitized patients with aplastic anemia, the allograft failure rate is low after conditioning with high dose CY alone. However, in patients sensitized by multiple blood transfusions, most allografts are rejected, and more intensive conditioning with a combination of TBI and CY is required to prevent take failure. Presently, there is not enough information regarding the numbers and drug sensitivity of memory lymphocytes in patients with advanced AID and how the size of these cell populations varies between patients.
We did not study fractionated TBI in EAE for two reasons. First, fractionation does not produce different effects from single dose TBI, provided the total dose is adjusted for the fractionation effect. Secondly, I did not anticipate the use of TBI in MS patients, as irradiation induced an acute exacerbation of the neurological symptoms in rats with EAE. This reaction recedes after 24 h, but was fatal in a small proportion of cases. It occurred even after a dose as low as 1.5 Gy. However, such adverse effects were not encountered so far in MS patients after treatment with high dose TBI and CY for conditioning [77 ].
ATG is used as part of the conditioning regimen in several clinical protocols of autologous BMT for severe AID [78 ] and juvenile idiopathic arthritis [79 ]. ATG was shown to protect against allogeneic GvHD in mice and monkeys even when administered before the BM [80 , 81 ]. If used for the conditioning in AID patients, it is recommended that the last dose be injected at least 24 h or less before the stem cell reinfusion. It may then act on residual lymphocytes in the recipient as well as on lymphocytes that are introduced with the graft [81 ].
Unfortunately, we could not evaluate the merits of ATG and its optimal application in conditioning for experimental AID, as the ant-rat ATG preparations available to us cross- reacted with hematopoietic stem cells. There remains an urgent need for more research on specific T lymphocytolytic agents to be applied in conditioning.
Postulated mechamisms of autologous stem cell transplantation and clinical relevance
The satisfactory, therapeutic results obtained with high dose lympomyeloablative conditioning regimens discussed above are easily understood if inflammatory AID are regarded as being initiated and maintained by activated T cells. Treatment of SLE and various connective tissue AID with moderate doses of cytoreductive drugs such as CY and methotrexate was already known to be effective in many cases so that the improved results with a higher dose is no surprise. The currently used conditioning for severe RA with 200 mg/kg CY has resulted in roughly 50% complete remissions and a high relapse rate. Apparently, this dose is not completely myeloablative, as it was associated with rapid hematological recovery when used without stem cell rescue for treatment of severe AID [82
]. In the latter study comprising 25 patients, the complete remission rate was also 50%, which suggests that the short-term responses at least are determined by the intensity of the conditioning and not by the cellular composition of the autografts.
Many clinical teams engaged in autologous BMT for AID so far prefer the use of CY alone or combined with anti-lymphocyte globulin (ALG). TBI is avoided mainly because of its several delayed side effects, e.g., the development of excess tumors, cataracts, and, in children, inhibition of skeletal growth. Yet, as pointed out in a recent review [83 ], the alkylating drug CY is also carcinogenic [84 , 85 ], as is prolonged immunosuppressive treatment [86 ]. Prolonged conventional treatment with high dose corticosteroids is known to cause growth inhibition and cataract, but a moderate dose of 4 Gy TBI does not have these sequels.
The experience with animal models indicates that there is room for improvement by intensifying the conditioning, but of course that would require hematological support with autologous stem cells unless more specific lymphocytolytic agents could be introduced. One promising drug is fludarabine. It was recently used successfully (120 mg/m2 over 4 days) in place of TBI in combination with CY and ATG for conditioning six transfusion-dependent patients with severe aplastic anemia for grafting of allogeneic BM [87 , 88 ]. There were no take failures, and all patients achieved full donor chimerism. In Cynomolgus monkeys, fludarabine (250 mg/m2 over 5 days) induced T- and B-cell lymphopenia and prolonged the survival of allogeneic skin grafts in naive and presensitized animals [89 ]. The drug schedule used caused transient neutropenia as the only side effect. Treatment of patients with refractory, severe RA with pulsed fludarabine induced a reduction of naive and memory CD4+ T cells [90 ]. High dose fludarabine (300 mg/m2 in two courses of 5 days) was used in combination with ALG followed by autologous stem cells in a pilot study for treating patients with various severe AID [91 ]. This regimen was not toxic, and the immediate responses resembled those after treatment with CY plus ATG, but follow-up was not long enough for other conclusions. There is certainly a strong need for sorting out, in the appropriate, established animal models, the advantages this drug has to offer and how it can be applied best.
The induction of remission and persistent tolerance
How are the excellent, curative results with autologous BM in experimental animals and the encouraging experience with many patients with various severe AID to be explained?
The most favored hypothesis is that the reconstitution of the immune system from a few stem cells represents a recapitulation of ontogenesis, which entails the acquisition of self-tolerance. Burt et al. [92 ] provided evidence for such a mechanism. They found persistence of T lymphocytes that react with fragments 68–82 of myelin basic protein (MBP) in the spinal cord of Lewis rats in spontaneous, clinical remission from acute EAE but not in rats that had been treated with high dose TBI and syngeneic BM to induce remission. Karussis et al. [69 ] claimed induction of tolerance in EAE mice after treatment with a 30% lethal dose of CY and rescue with syngeneic T cell-depleted BM [69 ]. The induction of EAE and rechallenge were with mouse spinal cord homogenate in CFA. They measured the proliferation response of lymphocytes obtained from lymph nodes to the antigens guinea pig MBP (GMBP) and tuberculin-purified protein derivative (PPD). The responses to GMBP were weak or negative before and after rechallenge in the treated animals as well as in the controls. The stimulation index with PPD increased from 8 before to 46 after rechallenge in the controls and from 2.3 to 3.8 in the treated group. It is doubtful if this can be regarded as evidence for tolerance. The spontaneous relapse rate was low in the treated group (one relapse in 15 mice) as compared with 21 relapses in 15 controls. Two out of eight mice suffered a relapse after rechallenge as compared with nine out of nine controls.
Apart from our ignorance about the details of tolerance at the cellular level, there is a clear tendency from the data in animals that suboptimal conditioning results in less complete responses (as in AA), more spontaneous relapses (as in EAE), and more induced relapses (as in AA and in EAE). Furthermore, following optimal conditioning, the induced relapse rate is higher with autologous or syngeneic stem cell grafts than with allogeneic transplants (Table 6) .
The causes of the different responses of AA and EAE to BMT remain subject to speculation. One attractive hypothesis is that if the genes determining susceptibility are weakly expressed, the resulting state of tolerance may not be broken easily. This may be the case in AA, where disease could not be reinduced after complete remission was obtained with autologous BMT. On the other hand, in EAE rats, relapses could be induced in a high proportion of animals after autologous BMT. At the extreme end of the range is the HLA-B27 transgenic rat that cannot even be brought into remission with syngeneic stem cells. These animals bear up to 150 copies of the B27 gene, which assumedly makes them respond to a large variety of environmental antigenic stimuli with autoimmune reactions, thereby precluding the development of autotolerance.
|
RECOMMENDATIONS |
|---|
|
TOPABSTRACTINTRODUCTIONRELEVANCE OF ANIMAL MODELS...ADJUVANT ARTHRITIS IN BUFFALO...EXPERIMENTAL AUTOIMMUNE...AID AS DISEASES OF...TREATMENT OF FULLY DEVELOPED...AUTOLOGOUS BMTEXPERIMENTAL DATA OF...RECOMMENDATIONSREFERENCES |
|---|
It stands to reason that reinfusing any numbers of T lymphocytes that would add substantially to the residual population should be avoided. A BM graft for an adult patient may contain as many as 4 x 109 and a mobilized peripheral blood cell graft, up to 2 x 1010 T lymphocytes as compared with an estimated 3 x 108 residual T lymphocytes. That estimate is based on conditioning regimes equivalent to 9 Gy TBI (single dose), which causes roughly a 3 log kill of lymphocytes. The total T cell population in an adult is taken as 3 x 1011. In view of the uncertainties of these estimates and in analogy with the policy of maximal purging of tumor cells in autologous BM grafts used for the treatment of leukemia, it was recommended to T cell-deplete the autograft as completely as current techniques allow. This strategy was supported by the relapse of all of the first five patients who received unmanipulated, autologous BM or mobilized peripheral blood cells [93 ]. The maximum number of reinfused T cells was recommended to be less than 105 per kg [94 ], requiring between 3 and 4 log depletion for BM and mobilized peripheral blood cells, respectively. In practice, 106 T cells/kg seems more realistic as a maximum; this number being well below the estimates of residual T cells. However, it is unknown which subpopulations of T lymphocytes are crucially involved in the development of relapses and what their radio- and chemosensitivity are.
Rescue with highly purified CD34+ stem cells is likely to cause an extended period of severe immunosuppression with increased risks of infections and lymphoproliferative malignancies. Several cases of fatal "activated macrophage syndrome" have occurred recently in children following treatment for juvenile chronic arthritis with highly purified autologous stem cells. It should be noted that the CD34+ cell selection techniques that are currently used also effectively remove B cells, natural killer cells, and macrophages, which may be unnecessary and possibly harmful. That has been the rationale for some clinical teams to change to purging methods that specifically deplete T lymphocytes only.
Finally, there is the option of using allogeneic stem cells, which in the EAE rat minimizes relapses and also showed a presumed graft-versus-autoimmune effect. Considering the higher risks of transplantation-associated mortality of allogeneic BMT, its exploration should be postponed until it becomes clear from the ongoing studies with autologous stem cells, which patients might benefit from allogeneic grafts. The most threatening side effect of allogeneic BMT is graft-versus-host disease, which represents an antiself immune reaction par excellence. Therefore, one risks replacing one severe disease with another. The use of allogeneic stem cells should specifically be discouraged in the treatment of SLE, systemic sclerosis, and related syndromes, as certain lesions induced by the graft-versus-host reactions, especially the ones associated with chronic GvHD, will be very hard to differentiate from lesions from a relapse of the original AID.
Received February 5, 2002; revised May 1, 2002; accepted June 4, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRELEVANCE OF ANIMAL MODELS...ADJUVANT ARTHRITIS IN BUFFALO...EXPERIMENTAL AUTOIMMUNE...AID AS DISEASES OF...TREATMENT OF FULLY DEVELOPED...AUTOLOGOUS BMTEXPERIMENTAL DATA OF...RECOMMENDATIONSREFERENCES |
|---|
Lewis and in Lewis F344 bone marrow chimeras Cell. Immunol. 136,278-290[Medline]
