A comparative study of reduced dose alemtuzumab in matched unrelated donor and related donor reduced intensity transplants

Summary In vivo T cell depletion with 100 mg alemtuzumab prevents graft‐versus‐host disease (GVHD) in reduced intensity conditioned transplants but is associated with delayed immune reconstitution, a higher risk of infection and relapse. De‐escalation studies have shown that a reduced dose of 30 mg is as effective as 100 mg in preventing GVHD in matched related donor (MRD) transplants. Dose reduction in matched unrelated donor (MUD) transplants is feasible but the comparative efficacy of alemtuzumab in this setting is not known and opinions vary widely concerning the optimal level of GVHD prophylaxis that should be achieved. Through retrospective analysis we made an objective comparison of MUD transplants receiving an empirically reduced dose of 60 mg, with MRD transplants receiving a 30 mg dose. We observed proportionate levels of alemtuzumab according to dose but an inverse relationship with body surface area particularly in MRD transplants. MUD transplants experienced more acute and chronic GVHD, higher T cell chimerism, more sustained use of ciclosporin and less need for donor lymphocyte infusion than MRD transplants. Thus, doubling the dose of alemtuzumab to 60 mg did not provide equivalent prevention of GVHD after MUD transplant although there was no difference in non‐relapse mortality or survival compared with MRD transplants.


Summary
In vivo T cell depletion with 100 mg alemtuzumab prevents graft-versushost disease (GVHD) in reduced intensity conditioned transplants but is associated with delayed immune reconstitution, a higher risk of infection and relapse. De-escalation studies have shown that a reduced dose of 30 mg is as effective as 100 mg in preventing GVHD in matched related donor (MRD) transplants. Dose reduction in matched unrelated donor (MUD) transplants is feasible but the comparative efficacy of alemtuzumab in this setting is not known and opinions vary widely concerning the optimal level of GVHD prophylaxis that should be achieved. Through retrospective analysis we made an objective comparison of MUD transplants receiving an empirically reduced dose of 60 mg, with MRD transplants receiving a 30 mg dose. We observed proportionate levels of alemtuzumab according to dose but an inverse relationship with body surface area particularly in MRD transplants. MUD transplants experienced more acute and chronic GVHD, higher T cell chimerism, more sustained use of ciclosporin and less need for donor lymphocyte infusion than MRD transplants. Thus, doubling the dose of alemtuzumab to 60 mg did not provide equivalent prevention of GVHD after MUD transplant although there was no difference in non-relapse mortality or survival compared with MRD transplants.
Alemtuzumab (CAMPATH 1H), a humanised immunoglobulin G1 (IgG1) monoclonal antibody against CD52, is a wellestablished agent for effecting in vivo T cell depletion in the setting of allogeneic haematopoietic stem cell transplantation. By reducing the number of both donor and recipient T cells, it allows engraftment, while reducing the incidence of acute and chronic graft-versus-host disease (GVHD) (Kottaridis et al, 2000;Chakraverty et al, 2002). Conventional alemtuzumab dosing (100 mg in five daily doses of 20 mg) has been shown to ablate acute and chronic GVHD almost completely in fludarabine and melphalan (FM) conditioned matched related donor (MRD) and matched unrelated donor (MUD) transplants (Kottaridis et al, 2000;Chakraverty et al, 2002). Subsequent studies have reported lower rates of chronic GVHD with equivalent overall survival compared with GVHD prophylaxis with methotrexate or mycophenolate mofetil (Perez-Simon et al, 2002;Delgado et al, 2008;van Besien et al, 2009;Malladi et al, 2009). This is achieved, however, at the expense of delayed immune reconstitution (Morris et al, 2003;Dodero et al, 2005), an increase in opportunistic infections (Chakrabarti et al, 2002a,b;Perez-Simon et al, 2002;Avivi et al, 2004) and higher incidence of mixed chimerism, which may impair graft-versus-leukaemia responses (Mackinnon et al, 1994).
The risks of T cell depletion can be ameliorated by donor lymphocyte infusions (DLIs) (Peggs et al, 2004(Peggs et al, , 2007Bloor et al, 2008;Thomson et al, 2010;Mohamedbhai et al, 2012;Liga et al, 2013) but there is a need to define minimum levels of T cell depletion that provide effective protection from GVHD in different settings. A formal dose de-escalation study in MRD transplants concluded that a single alemtuzumab dose of 30 mg on day À1 is sufficient to prevent GVHD   have been reported in mixed cohorts of MRD and MUD transplants for a wide range of alemtuzumab doses from 50 mg to as little as 10 mg (Faulkner et al, 2004;Khouri et al, 2004;Tholouli et al, 2008;Bertz et al, 2009;Spyridonidis et al, 2011;Gartner et al, 2013;Potter et al, 2014). These are difficult to interpret: first, they involve mixed cohorts of MRD and MUD transplants; and second, the definition of 'adequate' control of GHVD is entirely subjective with some centres tolerating much higher rates of acute GVHD and extensive chronic GVHD than others.
In the absence of a formal de-escalation study in MUD transplants and lack of rational criteria to define adequate GVHD prophylaxis, we compared an empirically reduced dose of 60 mg alemtuzumab in MUD transplants, with 30 mg for MRD transplants, previously defined by formal study . This offers an objective assessment of the potential of alemtuzumab to modify unrelated donor alloreactivity, relative to the level of a sibling donor.

Study design and ethics
The study included sequential patients transplanted with FM conditioning, following the introduction of de-escalated alemtuzumab dosing for MRD and MUD transplants in 2006. Clinical records were interrogated for GVHD, survival, chimerism testing and ciclosporin levels. All patients gave consent for clinical follow-up and post-transplant serum sampling for research purposes, according to protocols approved by the local research ethics committee of Northumberland and North Tyneside.

Conditioning regimen and alemtuzumab dosing
All patients received fludarabine 30 mg/m 2 per day from day À7 to À3 and melphalan 140 mg/m 2 day À2. Alemtuzumab 30 mg was given on day À2 for MRD transplants and on days À4 and À2 for MRD transplants (total 60 mg). Day À2 was preferred rather than day À1 because kinetic studies indicated a steep decline in levels in the first 24 h after infusion (Morris et al, 2003). We reasoned that the timing of stem cell infusion, which is not well controlled, would have less effect on the level of T cell depletion if we allowed 48 h following alemtuzumab adminstration, at which point alemtuzumab levels would be more stable. A minimum CD34 + stem cell dose of 4 9 10 6 /kg was infused in all patients. Historical patients treated with 100 mg of alemtuzumab received 20 mg daily from day À7 to À3, inclusive.

Post-transplant monitoring and management
Graft-versus-host disease prophylaxis was ciclosporin given at 2Á5 mg/kg twice daily from day À1 (with a target of 200-300 lg/l). Ciclosporin levels were monitored twice weekly by enzyme-linked immunosorbent assay. Tapering began at day +30 in the absence of GVHD. All patients received chemoprophylaxis against Pneumocystis jirovecii and Varicella Zoster. Serum cytomegalovirus (CMV) copy number was assessed twice weekly by polymerase chain reaction (PCR) and pre-emptive therapy was initiated if copy number rose above >10 4 . Peripheral blood and bone marrow chimerism was tested by short tandem repeat PCR on unfractionated bone marrow and magnetically-selected CD3 + and CD15 + fractions of peripheral blood. Monitoring took place monthly until day +100 and 3-monthly thereafter. A schedule of DLIs, escalating by half-log increments every 3 months was instituted for persistent partial chimerism, beginning at 3 9 10 5 /kg at 6 months. Patients with early relapse received 10-fold higher doses at 4-6 weekly intervals. Clinical assessment of acute GVHD was performed at least once weekly in the first 100 d using modified Glucksberg criteria (Hunt et al, 2001). Incidence of chronic GVHD was assessed at 1 year posttransplant and graded using both Seattle and National Institutes of Health criteria (Shulman et al, 1980;Filipovich et al, 2005). Body surface area was estimated by the method of Dubois (Dubois & Dubois, 1916).

Serum alemtuzumab measurement
Post-transplant serum samples were available from 20 MRD recipients and 13 MUD recipients at day +1 and 11 MRD recipients and nine MUD recipients at day 3-4. Alemtuzumab concentration was measured as previously described by a validated flow cytometry assay (Rebello & Hale, 2002).

Statistical methods
Intergroup analyses were performed with Student's t test for continuous variables, Mann-Whitney test for proportions and Chi-square for contingency. The three group-comparison of serum alemtuzumab levels used one-way analysis of variance (ANOVA). Multiple t-tests used for intergroup analysis of serum ciclosporin levels over time were subjected to sequential Bonferroni correction. Overall survival was analysed by the Kaplan Meier method with log-rank test to compare survival curves. Relapse and non-relapse mortality were analysed by the cumulative incidence method with relapse as a competing event for non-relapse mortality (NRM) and vice-versa. Cumulative incidence curves were compared with Gray's test. Hazard ratios were calculated by Cox Regression modelling. Graphing and statistical analyses were performed with GRAPHPAD PRISM version 6.0 (GraphPad Software, Inc. La Jolla, CA, USA) except for cumulative incidence, which was performed with R version 3.1.0 and Cox Regression, which was performed with SPSS version 21 (IBM United Kingdom Ltd., Portsmouth, UK). P < 0Á05 was recorded as significant.

Patient characteristics
From May 2006 to May 2009, 24 patients received MRD transplants and 19 received MUD transplants. Over this period a total of 51 patients were transplanted with reduced intensity conditioning (Fig 1). Eight were not included in the study because they were transplanted with different doses of alemtuzumab (two MRD and two MUD transplants) or the MUD transplants were less than 9/10 human leucocyte antigen (HLA)-matched (four patients). Patients included in the study were well matched for age, transplant indication and donor/recipient CMV status (Table I). All MRD recipients were 12/12 HLA-matched. Seventeen of the unrelated donors were 10/10 matched at high resolution. Two donors were 9/10 matched due to single allele mismatches in HLA-Cw.

Serum Alemtuzumab levels
Serum had been collected on days +1, +3-4 and +7 for immunomonitoring and a selection of samples was available for measurement of alemtuzumab levels. On day +1, the mean AE standard deviation (SD) serum alemtuzumab level was 2Á9 AE 1Á1 lg/ml in MRD transplants (30 mg) compared with 5Á4 AE 2Á2 lg/ml in MUD transplants (60 mg); (P = 0Á0002; Fig 2A). A collection of samples taken on day +3-4 was also available to compare dose-reduced transplants with historical patients receiving 100 mg alemtuzumab ( Fig 2B). The serum alemtuzumab levels for both MRD and MUD transplants were significantly lower with mean AE SD of 2Á3 AE 0Á8 and 4Á3 AE 1Á8 lg/ml, respectively, compared with 8Á2 AE 6Á0 lg/ml for recipients of 100 mg of alemtuzumab (P < 0Á0028). Alemtuzumab dosing regimens in adults do not usually take into account the size of the patient but an inverse relationship was observed between alemtuzumab level and body surface area (BSA) in both cohorts, reaching significance in MRD transplants (r 2 = 0Á70; P < 0Á0001; Fig 2C,D).

Discussion
The optimal dose of alemtuzumab for in vivo T cell depletion of reduced intensity conditioned transplantation is controversial and depends upon subjective judgments about the level of GVHD that tolerable or desirable, the perceived risk of relapse and the willingness of physicians to use pre-emptive DLI to bolster post-transplant immune reconstitution.
The overall survival has been reported to be comparable between patients receiving widely ranging levels of in vivo T cell depletion (Perez-Simon et al, 2002;Delgado et al, 2008;van Besien et al, 2009;Malladi et al, 2009). The use of alemtuzumab is generally associated with higher relapse risk but less non-relapse death and proponents argue that the quality of survivorship is improved by the prevention of extensive chronic GVHD (Perez-Simon et al, 2002), particularly in the T cell compartment.
Here we have evaluated 60 mg dosing for MUD transplants relative to the 30 mg dose, defined as optimal for stringent control of GVHD in MRD transplants . Monitoring of serum indicates that much lower levels are achieved than with 100 mg of alemtuzumab and that there is a predictable difference between 30 mg for MRD and 60 mg for MUD transplants, at day +1. We elected to give alemtuzumab 1 d earlier than Chakraverty et al (2010), at day À2, and the dose measured at day +1 was 30-40% lower than that reported on day 0 by these authors (2Á9 lg/ml for 30 mg and 5Á4 lg/ml for 60 mg in this study compared with approximately 5 and 7Á5 lg/ml, respectively). However, this did not appear to increase the incidence of acute GVHD grades II-IV in our MRD cohort (both studies <5%). We also demonstrated a significant inverse relationship with BSA for MRD transplants and a trend in the same direction for MUD transplants. Metresquared dosing is common in paediatric practise (Dodero et al, 2005;Shah et al, 2007) and it is possible that more predictable outcomes would be observed if a similar approach was taken in adults (Morris et al, 2003;Chakraverty et al, 2010). The use of one 30 mg dose (two 30 mg doses for MUD transplants) in our schedule also makes economic sense, as 30 mg is the presentation vial size of the drug. One possible drawback is that the infusional toxicity of alemtuzumab may be higher with a single dose than with several fractionated doses, although we did not observe major difficulties with this. Prior commencement of fludarabine (day À7) before the first dose of alemtuzumab (day À4) may have reduced the incidence of infusional toxicity in these patients.
Our study is limited by the retrospective nature and relatively small number of patients transplanted over the period of observation. A prospective study would have been preferred but reduced alemtuzumab dosing in MUD transplants was already common practise, making a large prospective study, similar to that performed in MRD transplants, difficult to conduct. Comparison with the sibling cohort receiving 30 mg of alemtuzumab, rather than the historical MUD cohort receiving 100 mg, was motivated by the fact that sibling 30 mg and MUD 60 mg could be compared contemporaneously and that the most carefully annotated dose regime in the literature is the 30 mg dose given to siblings. The power of the study to detect differences in survival is rather low as indicated by wide confidence intervals in the hazard ratios. However, our primary aim was to report outcomes related to the control of GVHD. The observation that overall survival remains relatively constant with different levels of GVHD prophylaxis has been reported in several other studies (Perez-Simon et al, 2002;Delgado et al, 2008;van Besien et al, 2009;Malladi et al, 2009).
Mixed cohorts of MRD and MUD transplant patients treated with lower doses of alemtuzumab have satisfactory outcomes overall (Faulkner et al, 2004;Khouri et al, 2004;Tholouli et al, 2008;Bertz et al, 2009;Spyridonidis et al, 2011;Gartner et al, 2013;Potter et al, 2014). In a large series of 58 MRD and 69 MUD transplants subjected to a reducing regimen of 40, 20 and 10 mg of alemtuzumab, it was possible to infer differences between the response of MRD and MUD patients to alemtuzumab (Bertz et al, 2009). For example, MUD transplants receiving 40 mg experienced 27% grade II-IV GVHD compared with MRD patients receiving 10 mg who only experienced 10% grade II-IV. This suggests that unrelated donors cause significantly more alloreactivity in this range of alemtuzumab dosing. In support of this, we found that 44% of MUD transplants experienced grade II-IV acute GVHD after 60 mg of alemtuzumab compared with 0% of MRD patients after 30 mg of alemtuzumab. Both reports are consistent in finding more acute and chronic GVHD with MUD transplants, even when alemtuzumab dosing was increased by several fold.
In keeping with higher rates of GVHD, we observed higher donor T cell chimerism at day 100 in MUD transplants and recorded lower use of DLI compared with MRD transplants managed with the same pre-emptive DLI policy. Higher indices of alloreactivity in MUD transplants were not due to more aggressive withdrawal of immunosuppression, as higher ciclosporin levels were maintained in these patients after day 50. Survival, relapse and non-relapse outcomes were comparable as in other studies. One limitation of this report is that we did not record quality of life, which may have been relatively impaired in MUD transplants owing to a higher rate of extensive chronic GVHD.
Our data suggest that 60 mg of alemtuzumab does not achieve equivalent control of GVHD compared with 30 mg in MRD transplants. However, it may be na€ ıve to assume that an 'equivalent' dose can be reached by simple titration. Review of MUD transplant patients receiving the original 100 mg dose indicates a 20% incidence of acute GVHD grades II-IV (Mead et al, 2010). Notably, this level of acute GVHD is still above that reported for MRD transplants with 30 mg dose (0-5%) and only marginally increases to 22% when mismatching is permitted in unrelated donors (Mead et al, 2010). These observations suggest that the relationship between alemtuzumab dose and antigenic disparity is nonlinear and that it may not be possible to achieve equivalent GVHD prophylaxis between MUD and MRD transplants simply by dose adjustment. Fortunately, as many studies illustrate, overall survival often remains stable despite significant variation in the burden of GVHD (Perez-Simon et al, 2002;Delgado et al, 2008;van Besien et al, 2009;Malladi et al, 2009). Further improvements in outcome may yet be achievable by paying close attention to individualised dosing of alemtuzumab based on BSA, expression of CD52 by malignant cells and the risk of graft rejection or relapse (Khouri et al, 2004;Chakraverty et al, 2010;Mead et al, 2010). Attention to long-term quality of life measurements is also paramount.