Cover Page
The handle http://hdl.handle.net/1887/47907 holds various files of this Leiden University dissertation.
Author: Torren, C.R. van der
Title: Investigating remission and relapse in type 1 diabetes. Immune correlates of clinical outcome in beta-cell replacement therapies
Issue Date: 2017-04-12
7
General Discussion
130 131
7
Chapter 7 General Discussion
7
Using the long-term benefit and risk as outcome in immune intervention studies is infeasible due to major costs and hinder on research progression. Using surrogate endpoints can shorten trials, if they reliably predict later benefit or risk [241]. Selecting measures for direct glucose control is tempting, since immediate improvement would be shown with long-term implications (Chapter 2). Near-normal HbA1c levels should be aimed for when treatment is commenced in all patients and is at that point relatively easy to achieve. Differences between treatment groups would therefore be minimal, which may preclude the treatment effect [204]. Higher insulin dose may reflect greater loss of endogenous insulin production, but is confounded by many factors.
Since glucose control is also important to prevent loss of beta-cell function and to prevent complications, minimizing insulin dose is controversial [3,204]. C-peptide is secreted by beta-cells together with insulin and can therefore be regarded as a correlate of beta-cell function. To assess maximal beta-cell potential, insulin and c-peptide secretion is often stimulated through high caloric intake or infusion of glucagon or arginine. C-peptide is not a direct measure of glucose control, but higher levels correlated with better glucose control and less long-term complications [149].
Efficacy
Side effects
aHSC-Tx
Cyclosporine
Rituximab Otelixizumab II Teplizumab II
GAD65
HSP60 Orale Insulin
Abatacept MMF + Daclizumab
Teplizumab III
Tregs Anti-IL1β
Otelixizumab III Azathioprine
IL-2+Rapamycin MTX
ATG
Insulin Vaccine
Pancreas Tx Islet Tx
insulin therap y
Complication resistant patients Complication
prone patients
Figure 7.1. Schematic representation of efficacy and side effects of immune intervention therapies for type 1 diabetes [1,12,14,33,39,57,60,62,72,76,83,98,108,114,115,123,138,139,158,160,164,165,16 7,178,198,199,209,217,258,260,265,270,301,307].
Curing type 1 diabetes requires functional beta-cells; they either may be protected at diagnosis, or thereafter need regeneration, reactivation or transplantation.
Ideally, type 1 diabetes development would be avoided by preventing onset of autoimmunity or intervention shortly after onset, however, this is not yet feasible [262]. Immune protection is essential in beta-cell centered treatment to prevent autoimmune destruction and rejection. In this thesis a critical appraisal is made of immune intervention strategies at onset of type 1 diabetes and immune challenges and intervention strategies were explored to transplanted beta-cells including those from alternative sources as treatment for type 1 diabetes. Thereby, providing insights that may support personalization and selection of immune suppression for beta-cell transplantation, and, further, uncovering immunological strengths and weaknesses of beta-cells from alternative sources.
POTENTIAL OF BETA-CELL PROTECTIVE THERAPY FOR TYPE 1 DIABETES
Immune intervention at diagnosis of type 1 diabetes with aggressive immune suppression and autologous stem cell transplantation can lead to longstanding insulin independence, however with risk of major side effects [60,307]. Further examination of intervention trials suggested a correlation between efficacy and side effects of treatment (Figure 7.1). For anti-CD3 T-cell therapy, dose reduction from the phase II to phase III clinical trials has prevented EBV reactivation, but has probably contributed to lack of efficacy of the phase III trials (Chapter 2) [9,12,114,134,139,258].
Disease specific immune modulatory therapies without immune suppression are likely to have less side effects. So far, however, their efficacy is also less robust. In Chapter 2, we discussed the possibility that coinciding unrelated immune challenges, e.g. viral infection or influenza vaccination, may influence efficacy of such immune modulation [160]. Such influences are hard to prove, let alone control, which may make these therapies unpredictable. Another problem is that immune modulation costs time in which beta-cell destruction can continue. Unequaled success of the autologous stem cell transplantation trial suggests early aggressive intervention may be required to be curative [60]. Novel insights that beta-cells, although not active, may persist in the pancreas years after diagnosis may relieve this urgency and create an opportunity for tolerance induction followed by reactivation of these beta- cells [59].
Treatment of type 1 diabetes with insulin brings major physical and psychological burden. Benefit of immune intervention therapies should therefore be considered in context of their influence on insulin therapy. Promoting insulin independence is a major improvement, especially when achieved permanently or with very slow decline of beta-cell function. DCCT/EDIC studies showed the benefit of temporary superior glucose control on complication risk, which still persists two decades after intervention [190]. Partial preservation of beta-cell function may already significantly improve long-term outcome, which is supported by a recent observational study in patients with long-standing diabetes, where remaining beta-cell function correlated with glucose control and negatively with complications [149].
7 7
outcome. We speculate that these growth factors influence transplantation outcome by enhancing beta-cell survival.
Serum markers studied in Chapter 4 are fundamental in diverse cell interactions and are highly interdependent. This complicates identification of a local response to beta- cells in serum levels of these markers. We employed clustering and linear modeling (limma) statistics to account for cytokine interdependency, however the statistical power of the small patient group studied was insufficient to determine effective discrimination on the complete data set. Individual cytokines that were identified included CCL3, which was previously identified as a predictor of simultaneous islet and kidney transplantation outcome [211]. This association proved not an independent predictor after adjustment for potential confounders, while IL-13 and IL-18 emerged as predictors. Studying (a selection of) the serum markers in a larger cohort may therefore identify additional biomarkers that we could not detect in our cohort.
We identified serum biomarkers in a small and relatively homogenous patient group.
Notably, variation was larger between patients than within a patient before versus after transplantation, suggesting the effect of immune suppression and improved glucose control on the homeostasis is relatively mild. Our ability to identify robust changes by transplantation is nonetheless demonstrated by confirmation of IL-7 increase after transplantation, which was previously reported to associate with homeostatic expansion of T-cell triggered by drug induced depletion [177]. Confirming this effect in another immune suppressive protocol also suggests that the T-cell expansion effect after depletion is universal to T-cell depleting induction therapy and supports the exploration of anti-IL-7 in islet transplantation once it comes available [91]. Which other identified serum markers will proof relevant in various beta-cell transplantation protocols remains to be determined.
HLA matching is a standard personalization of the graft for organ and stem cell transplantation, however this is infeasible for islet transplantation which often requires grafts from multiple donors. HLA matching can prevent graft rejection, but may promote recurrent autoimmunity through direct recognition of beta-cells by circulation memory cytotoxic T-cells [41]. Whether HLA matching is desirable becomes more relevant with increasing success of single donor islet transplantation.
Matching efficacy may vary per immune suppressive protocol, since alloreactive cytotoxic T-cells related to adverse graft outcome in sirolimus based immune suppressive protocols, but not after thymoglobulin, tacrolimus and mycophenolate mofetil immune suppression [234]. In the latter protocol, increased frequencies of autoreactive CD8 T-cells were identified, which we did not find when investigating samples after alemtuzumab induction therapy [302]. A case study of a patient who died after years of insulin independence with an islet graft showed preferential survival of the best matched islets [186]. In Chapter 3, we related donor specific islet graft infiltration and survival in the liver of a patient to results from peripheral immune monitoring. Unfortunately, we could not address relevance of HLA matching due to complete HLA class I mismatch of all donors to the patient.
C-peptide usually decreases in the first year after diagnosis and therefore seems the most appropriate outcome measure of early intervention trials [204]. For beta-cell transplantation trials all measures are expected to change robustly, while the effect of kidney function on c-peptide excretion needs to be taken into account in patients with kidney damage. Nonetheless, decline of c-peptide is likely to precede loss of insulin independence and glucose control and therefore has an important place in graft monitoring.
Immune protection is also needed for transplanted beta-cells. Transplantation of beta- cells is currently the only curative option for established type 1 diabetes. Continued immune suppression after transplantation is required to protect the graft from recurrent autoimmune responses and allograft rejection responses. This immune suppression comprises an important part of the side effects of this therapy. Together with donor shortage, these side effects limit application of beta-cell transplantation to patients with poorly controllable diabetes and those developing late complications.
Improving beta-cell transplant efficacy and reducing side effects can make this therapy suitable for more patients. Biomarkers to improve transplant success were explored in Chapter 4. Immune biomarkers can also help to guide choice of immune suppressive therapy (Chapter 4). In Chapter 5 possibilities to prevent side effects through reducing immune suppression and preventing alloreactive sensitization were investigated. Alternative sources of beta-cells may overcome donor shortage, but the required immune suppression may depend on cell origin, e.g. autologous, allogeneic or xenogeneic, and conferred immune privilege (Chapter 6). Encapsulation of transplanted beta-cells might replace immune suppression in the future and would additionally allow retrieval of beta-cells after transplantation if macro-capsules are used.
PERSONALIZED CARE TO OPTIMIZE BETA-CELL SURVIVAL
Personalization of intervention therapies on patient and disease characteristics may improve outcome for patients and increase treatment efficacy. In Chapter 2, we identified patient age and ethnicity as important covariates from the subgroup analysis of immune intervention trials after diagnosis. Age and ethnicity are crude measures, which probably reflect underlying variation in disease characteristic and drug dynamics and kinetics [316]. Unraveling these patient-to-patient differences and more specific biomarkers may better guide patient and treatment selection [6,241,316].
For islet transplantation, autoimmune and alloreactive responses through T-cells and antibodies were shown to predict and/or reflect outcome [44,118,124,127,212,271].
Blood is rich in serum proteins involved in intercellular communication, which may therefore reflect the metabolic and immunological state of the body. In Chapter 4 we studied a large panel of serum proteins before and 1 year after islet transplantation and identified profiles that correlate with transplantation outcome. Only few of these serum proteins, all cytokines and chemokines, have been studied in human islet cultures, after islet alone or islet after kidney transplantation [49,171,177,211]. Our serum protein profiles also identified diverse growth factors to be associated with
134 135
7
Chapter 7 General Discussion
7
patient who died several months after transplantation (Chapter 3). The only study describing in situ studies and immune monitoring did not identify any autoimmune or rejection responses at the graft or in blood [264]. This patient developed increasing alloreactive T-cells after transplantation and all identified islets had lymphocytes at the border including many cytotoxic T-cells. A donor cell free margin around most of these infiltrating cells suggests the islets were under immune attack.
Further determination of immune infiltrate specificity remains challenging. We identified islets from half of the donors which included donors from the first and second transplantation. This suggests multiple donors still contributed to graft function. Due to limitation of our islet donor identification method to cryopreserved tissue, too few islets could be identified to allow investigating preferential destruction or survival of islets from specific donors. Upcoming alloreactivity to all donors in peripheral blood monitoring matched with infiltration of all identified islets. Confirmation of T-cell specificity in biopsies has been done before using tetramers [59]. For tetramer studies, knowledge of potentially targeted HLA-peptide complexes is required. In this case, lack of matching HLA between any donor and the recipient precluded these studies. Lack of donor and recipient HLA match also seemed to exclude infiltration by autoreactive CTL. In concordance, the islets showed no indication of selective beta- cell loss. Our data support the relevance of alloreactive T-cell responses measured in peripheral blood.
To identify donor origin of individual islets while allowing further histological assessment, we optimized an HLA specific staining using human alloreactive antibodies (Chapter 3). We observed large variation in HLA staining intensity and autofluorescence between tissues from various organs, complicating identification of human islets with low staining intensity transplanted to liver tissue with high autofluorescence. Enhanced indirect immune fluorescence allowed consensus of identification of donors of various islets by three observers. Separate training and blinded scoring would have improved credibility, but was not feasible with the few available slides. Increased HLA staining intensity of part of the beta-cells facilitated islet donor identification. This variation in staining intensity may have been caused by HLA hyperexpression on cells that were involved in or near an immune response.
HLA hyperexpression by islets has previously been described in islets with insulitis in patients with type 1 diabetes [59]. In Chapter 6, we also showed increased HLA expression and alloreactive antibody targeting on alternative beta-cells after stimulation with inflammatory cytokines. These studies indicate HLA hyperexpression can be important in beta-cell transplantation and show that beta-cell can be targeted by human alloreactive antibodies. Preventing HLA hyperexpression through cytokine suppression or innate immune suppression may therefore improve transplantation outcome.
Some recent trials included innate immune suppression in their protocol with varying success [23,49,80,161,220,228,278]. Innate immune responses pave the way for adaptive immunity, for example by producing cytokines that upregulate HLA on beta-cells, and can negatively affect islet engraftment [50]. A recent study showed efficacy of innate immune suppression through CXCR1/2 blocking to improve islet Immune biomarkers may also help to personalize immune suppressive therapy
after transplantation (Chapter 4). We showed that both thymoglobulin and daclizumab induction therapy have similar efficacy in protection from acute kidney rejection episodes after simultaneous kidney and pancreas transplantation.
However, daclizumab may be preferred, since more side effects have been reported for thymoglobulin including reactivation of CMV viremia in this cohort [7,37,110,125,173,183,197,306]. We identified that pretransplant seropositivity for GAD antibody predicts more rejection episodes under daclizumab, but not under thymoglobulin induction therapy. Therefore, patients with GAD antibodies may be better off when treated with thymoglobulin induction therapy.
Studies on autoimmune antibodies so far identified a post-transplant rise in autoantibody titer as indicator of graft loss, rather than pretransplant seropositivity.
Also, these studies identified ZnT8 and IA-2, rather than GAD, antibodies as most predictive of pancreas graft loss [31,35,67,169,196,212,232,304]. The specificity of GAD autoantibodies and not IA-2 autoantibodies to predict kidney rejection episodes after daclizumab induction therapy suggests a kidney specific response, rather than a reflection of upcoming auto- and alloreactive immune responses to both grafts.
Indeed, GAD antigen is expressed in the kidney to which cross-reactivity may arise [40,290]. The relevance of GAD antibodies as predictor of rejection episodes may therefore be independent of the pancreas graft and extend to all kidney transplantations to patients with type 1 diabetes.
Immune biomarkers identified to date have not been implemented in beta-cell transplantation programs for several reasons. Their discriminative power or statistical validation may fall short, exemplified by serum cytokines identified in Chapter 4. The biomarkers may relate to rarely used or outdated treatment, e.g. for absence of GAD antibodies advocating the use of daclizumab induction therapy (Chapter 6). Inversely, immune suppressive protocols are under continuous development including new agents for which biomarkers remain to be identified and validated. The biomarker may not be sufficiently related in time or insufficiently practical to measure to use for intervention, e.g. monitoring antibody increase to predict graft loss. Last but not least, the predicted outcome measure may not be relevant enough in clinical practice, e.g. T-cell autoimmunity can identify patients that reach insulin independence after islet transplantation, however exclusion of patients not predicted to achieve insulin independence is unethical since they can benefit greatly nonetheless.
UNDERSTANDING AND IMPROVING IMMUNE SUPPRESSION
Better understanding of immune responses after transplantation may help to optimize immune suppressive therapy. Direct investigation of islets transplanted to the liver is rarely possible, since investigating islets through regular needle biopsies proved infeasible and taking large liver biopsies is too invasive [287]. Therefore, auto- and allograft specific immune responses have been studied in peripheral blood of patients receiving islet transplants and related to transplant success [124,127,212,234].
We investigated whether these immune responses measured in blood related to local immune infiltration or destruction of a beta-cell graft in liver sections of a
7 7
an effective regulatory response [66,176,317]. Further, the choice of induction immunotherapy at transplantation may hinder tolerance. Induction with anti–IL-2R antibody or sirolimus resulted in homeostatic expansion of remaining T-cells, including autoimmune T-cells [177]. Similar responses may be seen after thymoglobulin, which was used in our patients, although T-cell numbers will be suppressed longer [118]. Alemtuzumab treatment also results in prolonged lymphopenia, but in these patients we identified a unique IL-10 producing phenotype. This could be the first step towards operational tolerance.
Immune protective capsules may circumvent the need for immune suppression altogether. These encapsulated transplants would require sufficient space as well as early vascularization to ensure oxygen and nutrients for beta-cell survival. We observed a peak in T-cell autoimmunity in a patient after intra-abdominal transplantation with alginate microencapsulated islets (unpublished data). At closer examination, several alginate capsules had been disintegrated and therefore left beta-cells exposed to the immune system, while other capsules could be recovered and contained living beta-cells (personal communication, Bart Keymeulen, Brussels, Belgium). The encapsulated islets appeared to have been protected from autoreactive responses in the patient. Islets released from the disintegrated capsules may have provoked the measured immune responses. Anticipating on transplantation of beta-cells from alternative sources with a potentially higher teratogenic and tumorigenic risk, these immune responses may be regarded as safety biomarkers or even a safety mechanisms to eliminate cells escaping from the capsule. Nonetheless, immune responses provoked by escaping cells may induce alloreactive sensitization, which can threaten future transplants.
ANTICIPATING ON LATE SEQUELLAE
Late side effects may arise after islet graft function is lost and immune suppression is discontinued. These sequellae should be avoided if possible and, at least, be taken into account when planning beta-cell transplantation. One such risk is developing Graves (autoimmune thyroid) disease, which was shown to be predicted by the presence of thyroid peroxidase antibodies before transplantation [95]. Another risk is developing alloreactive antibodies, which may induce acute rejection of future transplants [51,281]. For patients with diabetes this is particularly relevant, since they may require kidney transplantation if end-stage diabetic nephropathy develops.
Our in situ investigation of transplanted islet also confirmed these antibodies can bind transplanted beta-cells and thus may be a challenge for repeated islet transplantations (Chapter 3).
Cytotoxic alloreactive antibodies may arise after islet transplantation in 30% of patients on immune suppression and may further increase up to 58% if immune suppression is discontinued [45,46,193]. In Chapter 5, we studied alloreactive antibodies after graft failure and discontinuation of immune suppression. In this cohort, islets were used after several days of culture. This increases beta-cell purity at cost of cell yield, which increases the number of donor required for successful transplantation [135]. HLA matching is not feasible in islet transplantation, therefore more donors transplantation outcome [49]. Also, broad immune suppression through anti-CD52
induction therapy (alemtuzumab) recently combined with other anti-inflammatory drugs greatly improved outcome with more than 50% insulin independence after 5 years (personal communication, James Shapiro, Edmonton, Canada). Preliminary data of post-transplant samples of these patients showed greatly suppressed immune responses during two years after alemtuzumab induction therapy. Remaining immune responses showed a unique IL-10 producing phenotype (unpublished data).
This supports broad and prolonged immune suppression, which favored outcome and may have allowed a tolerant immune response supporting long-term effect.
REDUCING SIDE EFFECTS OF IMMUNE SUPPRESSION
The greatest side effects from beta-cell transplantation come from lifelong immune suppression, which may lead to opportunistic infections, post-transplant lymphoproliferative disease and increased risk of malignancy [87,207,279]. Reducing immune suppression to minimal doses would decrease these risks. Optimally, graft tolerance is induced and no further immune suppression is required. Operational graft tolerance has been described in liver and kidney transplantation [26,77].
Transplanted livers can convey allograft tolerance, which indicates the liver may be immune privileged [26]. We investigated 5 patients receiving an islet transplant to the liver in whom maintenance immune suppression was tapered more than one year after transplantation (Chapter 5). Unfortunately, beta-cell function declined in all patients during tapering of immune suppression. Rather than part of general graft decline after transplantation which can be multifactorial [245], steep deterioration of c-peptide production suggested loss of graft function may be immune mediated.
Monitoring these patients for auto- and alloimmunity during tapering of immune suppression showed diverse auto- and alloreactive responses, which may have caused graft destruction. Further analysis showed the avidity of alloreactive cytotoxic T-cells increased during tapering of immune suppression and loss of graft function, rather than their frequency, which could be reversed in vitro when immune suppressive therapy was resumed. This implies that the quality of the reaction may be more important than the actual frequencies of alloreactive CTL, which was previously noted in heart and corneal transplantation [202,235]. Suppression of T-cells remained effective when restoring immune suppression levels in vitro, which suggests resuming immune suppression to initial levels in patients may abrogate upcoming immune responses during attempts to taper immune suppression.
Our results showed no indication for induced operational tolerance. In vitro addition of tacrolimus to autoreactive T-cell responses shifted cytokine production from pro- inflammatory (IFNγ, IL-13) to anti-inflammatory (IL-10) dominated. Therefore, a form of drug mediated tolerance may exist during treatment with tacrolimus. In that case, the drug mediated tolerance does not seem to convey to operational tolerance by regulatory immune cells. Regulatory T-cells may again be opposed by tacrolimus through hampering of IL-2 signalling, which is an important cytokine for them [208].
Whether mycophenolate mofetil, which was not tapered during the study, positively or negatively affects regulatory T-cells remains unclear, but it may have prevented
138 139
7
Chapter 7 General Discussion
7
success through transplantation of optimized beta-cell dose [135]. Each alternative source of beta-cells has its own traits which may include differences in immune responses. Therefore each alternative beta-cell type may require tailored immune protection (Table 7.1).
Adaptive immune responses that threaten alternative beta-cells depend on their HLA overlap with the recipient which may be influenced by beta-cell ancestry. Beta-cells derived from a patient’s own cells, e.g. through induced pluripotent stem cells, will have complete HLA compatibility which prevents alloreactivity at risk of recurrent autoimmune responses. Animal beta-cells lack human HLA and are therefore no target for auto- or alloreactive T-cells, but may be overwhelmed by innate and animal antigen specific antibody responses instead [58]. In Chapter 6, we showed that allogeneic beta-cells of alternative sources have varying sensitivity to auto- and alloreactive responses that currently threaten islet and pancreas allografts.
Both embryonic stem cells (hESC) derived pancreatic endoderm and beta-cell lines from fetal origin express low levels of HLA which protected them from cytotoxic T-cells and alloreactive antibody mediated destruction (Chapter 6). This is an immune evasion mechanism known to be exploited by stem cells [79]. Differentiation of hESC derived pancreatic endoderm to endocrine cells increased HLA expression and vulnerability to cytotoxic T-cells and alloreactive antibodies. Similar increase in HLA expression and immune sensitivity could be induced by exposure to inflammatory cytokines. These cytokines may be produced by innate immune responses and by recurrent autoimmunity. We established that human beta-cell specific T-cell response supernatant effectively induced HLA upregulation, similarly inducing immune vulnerability to alloreactive antibodies and cytotoxic T-cells. These limitations to the immune privilege of these alternative beta-cells needs to be taken into account when designing transplantation trials.
Inflammatory cytokines can also directly lead to apoptosis of hESC derived beta- cells and beta-cells lines. Cytokine induced beta-cell death has been hypothesized to contribute to type 1 diabetes in inflamed islets [11,74,221]. However, cytokine induced apoptosis was not beta-cell specific on cadaveric islets nor on hESC derived endocrine cells and their progenitors (unpublished data) [221]. Nonetheless, inflammatory cytokines adversely affect these alternative beta-cells.
An innate immune response is inevitably induced by any transplant procedure and includes granulocyte and NK-cell infiltration, complement responses and production of aforementioned inflammatory cytokines. Expression of cell surface markers protect cells from innate immune destruction. Against complement destruction, beta- cell lines and hESC derived pancreatic progenitor and endocrine cells expressed inhibitory receptors (Chapter 6). However, NK-cell attack is partially suppressed by HLA expression, which was low on beta-cell lines and hESC beta-cell progenitors.
Our results on beta-cell lines showed that the inflammation dependent upregulation of HLA could help protect these cells from NK-cells. Actual NK-cell destruction may therefore be foiled, however, at price of exposure to adaptive immunity which may be detrimental in pre-immunized patients.
mean more HLA mismatches and more targets for alloreactive antibodies to arise against. Despite a median of six donors per islet recipient, we showed that the risk of developing cytotoxic HLA antibodies was limited (6%). In another cohort, it was also recorded that the number of donors is unrelated to HLA antibody formation in islet transplantation [193]. Although the immune suppression protocols used were slightly different from previously reported cohorts, islet culture seems to be the main contributing factor to the lower cytotoxic alloreactive antibody sensitization.
We further explored risk factors for alloreactive antibody development based on antibodies detected by Luminex technology. This technology is more sensitive to detect alloreactive antibodies, but the relevance of these antibodies is disputed [52].
Choice of immune suppression did not influence alloreactive antibody formation, except for steroid use. This is in concordance with literature [81]. However, steroids have a negative effect on beta-cell function and are therefore not preferred for islet transplantation.
Beta-cell purity was the most important factor to prevent HLA class I and II antibodies detected by Luminex. This suggests alloreactive antibody formation is driven by non- beta-cells in the transplant. For HLA class I this may be explained by low expression on beta-cells compared to other cells, which we also noted in Chapter 3 and on alternative source beta-cell like cells in Chapter 6. HLA class II is per definition not expressed on beta-cells, but may be in the graft through endothelial cells, dendritic cells and duct cells [55,85,206].
The absence of a relation between alloreactive antibody formation and number of HLA differences was further substantiated by lack of correlation with HLA amino acid triplet mismatches. Triplet mismatches take into account the overlap between various HLA molecules, which are therefore not immune reactive, and have been shown to predict graft rejection in well HLA matched kidney transplantation [61].
A possible explanation for this difference is that small HLA differences between donor and recipient allow rejection responses, while this effect quickly reaches a plateau with major HLA differences. Another explanation would be that many of these antibodies arise as bystander effect of an inflammatory response, which could also be autoimmune for islets transplanted to patients with type 1 diabetes. This is supported by the observation that most alloreactive antibodies identified by Luminex are not specific for donor HLA, but for (third-party) HLA which was not involved in the transplantation. The overrepresentation of women with previous pregnancy in our cohort suggests these antibodies may come from memory B-cells and could be directed to HLA which their offspring inherited from their father. In line with this, Luminex detected antibodies have been identified in 45% of women after pregnancy [119]. The induction of these antibodies, which are not donor specific and are not directly cytotoxic, does not seem to require additional preventive measures.
POTENTIAL OF ALTERNATIVE BETA-CELLS
Beta-cells from alternative sources can overcome shortage of donor organs thus increasing availability of beta-cells for research and treatment. This would allow broader application of beta-cell transplantation and potentially improving transplant
7 7
Beta-cell lines and hESC derived beta-cells vary in their potential for research and transplantation. The introduction of oncogenes in the beta-cell lines preclude their use for clinical transplantation until these genes can be neutralized with great certainty.
In contrast, a phase 1/2 clinical study has recently started for the hESC derived beta-cells to assess their safety and efficacy (clinicaltrials.gov, NCT02239354).
In contrast, research experience with these cells favors the beta-cell lines, which were relatively easy to maintain and provided a stable and homogeneous target population for immunological assays. They further allowed to study the response of beta-cells to immune attack separate from other pancreatic endocrine cells, which remains impossible for many assays in mixed cell populations. Conversely, hESC derived beta-cells allowed study of beta-cells in an endocrine cell population context, which better reflects the study of cadaveric human islets. Both cell lines differ from adult human beta-cells by their manipulated growth starting from embryonic or fetal stage to become functional beta-cells. How immunological characteristics differ from adult beta-cells needs to be determined, but variation in HLA expression and complement mediated attack sensitivity between the alternative beta-cells argues relevant differences exist. Therefore, extrapolation of results from both alternative beta-cells to adult human beta-cells needs to be considered with care.
FUTURE DIRECTIONS AND CONCLUDING REMARKS
Beta-cell protection has significantly improved over the past decade resulting in prolonged remission of the first patients with type 1 diabetes through immune therapy after diagnosis and improving results of beta-cell transplantation. However, for extensive application to supplement or replace insulin therapy further improvement is required. We observed that individual differences may have greatly attributed to treatment efficacy and that immune biomarkers can help determine these differences.
Personalization of immune therapy could therefore swing the balance to efficacy over side effects. The data in this thesis also indicate the risk of HLA (hyper)expression, which may be countered by addition of innate immune suppression to the current protocols for islet transplantation. Tolerance induction after islet transplantation may greatly reduce long-term immune side effects. Our study did not show tolerance under the studied immune suppression. Strict immune monitoring of other immune suppressive and tapering protocols may identify more successful therapies. Cellular tolerance induction protocols currently under investigation after type 1 diabetes onset may also be applicable after islet transplantation. Alternative beta-cells can overcome donor shortage which may extend indications for beta-cell transplantation. Human embryonic stem cell derived beta-cells now require clinical evaluation for safety and efficacy. Our results show that monitoring auto- and alloreactive immune responses may be equally relevant for these trials as for current beta-cell transplantation. If immune protective encapsulation of beta-cells is achieved and proven successful for transplantation, this also would be a major step forward since it could overcome the side effects of immune suppression. Recent progress and new opportunities of beta- cell protection and replacement incite hope that we can ultimately induce lasting remission in patients with type 1 diabetes.
PancreasIsletsPigsBeta-cell linehESCiPS cells*Trans- differentiation from other cell types*
Innate immune response
Shielded by organNormalAggressiveNormalNormalNormalNormal
Autoimmune response
HLA dependentHLA dependentNoneProtected by low HLAInitially protected
Potentially high if autologous origin Potentially high if autologous origin
Alloreactive response HLA dependentHLA dependentNoneProtected by low HLAInitially protectedOnly if allogeneic originOnly if allogeneic origin
Immune suppression target
T-cellsT-cells
Antibodies/ complement
Inflammation
Inflammation, later
T-cellsT-cells/ inflammation?T-cells/ inflammation? OpportunitiesTailored ISTailored IS,
alternatives sites, single donor
Tx, encapsulation
Hypoimmunogenic islets, encapsulation
Excisable transgenes
Clinical
implementation, encapsulated Autologous cells, HLA-matching
Increase donor organ yield Table 7.1. Immunity to beta-cells of different sources. *No experimental data on human beta-cells available. hESC: human embryonic stem cells; iPS cells: induced pluripotent stem cells; HLA: human leukocyte antigen; IS: immunosuppression.
