Patients and methods

Since 1996 we have performed 75 allogeneic transplants for lymphoid malignancy (excluding ALL) of whom 17 patients have received DLI for either resistant disease (n=7) or disease relapse (n=10). Patients were eligible for DLI if they had radiological or biopsy evidence of persisting disease post transplant or if they had relapsed disease. Patients had a median age of 51 years (30–62 years) and 15 had undergone a transplant from an HLA-matched sibling donor (n=13) or 1 antigen mismatched sibling donor (n=2). In two cases, the donor was a matched unrelated donor. The initial diagnoses were diffuse large B-cell lymphoma (DLCL, n=4) mantle cell lymphoma (MCL) (n=4), follicular NHL (n=4) and CLL (n=5). The latter group included one patient with CLL at diagnosis (patient 11), who had undergone Richter's transformation at relapse and was therefore treated as aggressive disease (Table 1). The conditioning regimen at time of transplant was BEAM-campath (n=11),⁹ fludarabine, melphalan and campath (n=4) or TBI (12 Gy in 6 fractions) and cyclophosphamide (n=2). Of the 17 patients, 15 had received in vivo campath IG or campath 1H (Alemtuzumab) given prior to transplant (50 mg over days -5 to -1).

Table 1 - Patient demographics.

Full table

The median time from transplantation to DLI infusion was 10 months (range 3–54 months) and was 6 months for patients receiving DLI for residual disease compared to 12 months for patients receiving DLI for disease relapse. No patient had active GVHD at the time of DLI therapy. At the time of DLI therapy seven patients had full donor chimerism, four patients had mixed chimerism and in six patients the chimerism status was not performed. Chimerism was performed on DNA extracted from whole blood with the identification of an informative microsatellite marker as previously described.¹⁰

Schedule of DLI administration

Of the 17 patients, nine received chemotherapy prior to DLI infusion. This included all patients with a diagnosis of DLCL and mantle cell NHL as well as the patient with Richter's transformation of CLL. Patients considered to have low-grade disease (CLL or follicular lymphoma) received DLI alone or in one case DLI following local radiotherapy for lymphadenopathy in the right groin.

Chemotherapy prior to DLI was with 2–3 cycles of CHOP+/- rituximab (n=2), IVE (ifosfamide, etoposide, epirubicin +/- rituximab) (n=2), FLUDAP (fludarabine, Ara-C, dexamethasone, cisplatinum) (n=2), FCR (fludarabine, cyclophosphamide, rituximab) (n=2), or vinblastine/dexamethasone (n=1) (Table 1). The choice of pre-DLI chemotherapy depended upon the therapy that the individual patient had received pre-transplant. For patients receiving pre-DLI chemotherapy, DLI were given following maximum response.

Of the nine patients who received chemotherapy prior to DLI, four had chemosensitive disease and five had chemoresistant disease; however, none of the chemosensitive patients were in CR at the time of DLI infusion. DLI were given in an escalating protocol at a minimum of 3 monthly intervals in the absence of a disease response to the previous DLI. The median initial cell dose given from sibling donors was 2 10⁷ CD3+ cells/kg (range 0.5–5 10⁷/kg). Eight patients received a second infusion with a median of 5 10⁷/kg CD3+ cells (range 5–10 10^7/kg) and one patient received a 3rd infusion comprising 5 10⁷/kg CD3+cells. The two patients receiving DLI from unrelated donors received 0.2 and 0.5 10⁷ CD3+ cells/kg at course 1; 0.5 and 1 10⁷ CD3 cells/kg at course 2 and one patient received 1.5 10⁷ CD3 cells/kg at course 3. Full details of the DLI schedule for each patient including pre-DLI therapy is given in Tables 1 and 2. The choice of initial DLI dose depended upon the donor type (sibling or unrelated), the time of relapse post transplant and the grade of disease. Survival estimates were calculated by the method of Kaplan and Meier. Overall survival (OS) was calculated from the time of the first DLI until death from any cause. Progression-free survival (PFS) was calculated from DLI until progression or death from any cause. The minimum follow-up of surviving patients was 12 months. Acute GVHD was graded 0 to IV according to the standard criteria.¹¹ Chronic GVHD was assessed in patients surviving more than 90 days after DLI and was defined as limited or extensive.¹²

Table 2 - Response to DLI and outcomes.

Full table

Results

Response to DLI-based therapy

The response to each cycle of DLI therapy is shown as a flow diagram in Figure 1. Following the first infusion of DLI, five patients were in CR and received no further therapy; eight patients had persisting disease without disease progression and received a second infusion of DLI; three patients had disease progression and received no further therapy. One patient (number 16) with MCL had initial disease progression immediately post-DLI with progression of abdominal lymphadenopathy and received pulsed dexamethasone and vinblastine with abdominal radiotherapy and later developed extensive GVHD followed by achievement of CR.

Figure 1.

Flow diagram of response to DLI therapy.

Full figure and legend (13K)

Of the eight patients who received a second DLI infusion two achieved a CR and three patients had disease progression receiving no further therapy. One of these patients (number 7) developed extensive chronic GVHD but with biopsy-proven persisting liver lymphoma, which remained stable for 15 months post-DLI before progression. Another patient (number 17) developed graft failure and eventually received a second allogeneic transplant from the same donor with low-dose TBI (2 Gy) and fludarabine conditioning, the patient achieved a CR following the development of extensive chronic GVHD and remains in CR at 24 months post transplant. Two patients with persisting disease received a 3rd DLI infusion and both patients achieved a CR. Therefore, a total of 11 patients finally achieved CR (64%), this included both patients with DLI from an unrelated donor. The median cell dose required to achieve CR for siblings was 2 10⁷/kg CD3+ cells (range 0.5–2.5 10⁷/kg), whereas the median cell dose given to the nonresponders was 5 10⁷/kg CD3+cells (range 2–10 10⁷/kg).

When analysed by disease group, patients with low-grade disease had a high response rate. Thus, CR was achieved in three out of four patients with follicular lymphoma and three out of four patients with CLL. CR was also achieved in all five patients with MCL. In contrast none of the five patients with DLCL/Richters transformation of CLL achieved CR. Response was achieved in all seven patients with full donor chimerism and in two out of four with mixed chimerism.

Toxicity, relapse and survival

Acute GVHD developed in seven patients out of 16 who survived >3 months post-DLI. Four patients developed grade II GVHD and three developed grade III/IV disease requiring treatment in all cases. One patient with MCL who received 1 10⁷/kg CD3+ cells at 11 months with no response and then 5 10⁷/kg cells at 14 months died of grade IV acute GVHD in CR. There were no other deaths in CR. Chronic GVHD developed in eight of nine patients who achieved CR post-DLI. One patient who had mixed chimerism at relapse (74% donor) developed pancytopenia following the second DLI with a complete loss of donor graft function and then received a second transplant from the same donor with good engraftment.

The risk of relapse is low and only one patient with MCL has relapsed at 18 months post-DLI, despite ongoing chronic GVHD and has received no further DLI therapy. No other patient has relapsed with a median follow-up of 34 months (range12–64) post-DLI. Overall seven patients have died, six from disease progression and one from GVHD. Overall survival is currently estimated as 58% (Figure 2) and PFS as 52% at 3 years.

Figure 2.

Overall survival following DLI therapy.

Full figure and legend (30K)

Discussion

Allogeneic HSCT is being increasingly used in patients with resistant and relapsed lymphoid malignancy particularly using RIC conditioning.² Preliminary results suggest that RIC conditioning can result in a high response rate and a low risk of disease relapse for patients with low-grade lymphoma, particularly for follicular lymphoma and also for CLL.¹⁰ Promising results have also been reported in peripheral T-cell lymphoma¹³ and also in MCL particularly when using T-replete transplants.^{14, 15} In vivo T-cell depletion, particularly using variable doses and scheduling of Alemtuzumab have been used in some series of RIC transplants in lymphoma^{10, 7} and although this approach is highly effective in limiting both morbidity and mortality from GVHD^{9, 16} there is the potential for a greater risk of disease relapse because of an abrogation of GVL activity. This may be particularly true of mantle cell and aggressive NHL.¹⁰ The use of DLI is therefore an integral part of any RIC regimen using T-cell depletion. Here we have evaluated the response to DLI in a series of 17 patients with either residual disease or relapse following allogeneic HSCT for lymphoid malignancy, the majority of whom had received pre-transplant Campath. Our strategy was to use DLI alone for patients with low-grade disease including follicular NHL and CLL where the pace of disease relapse was slow enough to permit sufficient time to permit the development of a GVL effect. For patients with aggressive disease or MCL, we used initial cytoreductive treatment in an attempt to prevent rapid tumour growth from overcoming any GVL effect. We used an escalating schedule of DLI starting between 0.5 and 5.0 10⁷/kg CD3+ cells depending upon disease subtype and response to pre-DLI chemotherapy. Patients receiving DLI from unrelated donors received lower doses of cells.

Our results show an overall high response rate to this approach with 10 out of 17 patients achieving a CR following one to three DLI infusions. A further patient also achieved a CR following a second transplant from the same donor following graft rejection post-DLI giving a final overall CR rate of 64 %. The major finding of our study was that the response to DLI was strongly related to the histological subtype of disease. Patients with follicular lymphoma, CLL and MCL have a high response rate to DLI therapy. Overall 10 of the 12 patients with these diagnoses achieved a CR (83%). The high response rate in MCL may in part be related to our strategy of using pre-DLI chemotherapy, although no patient was in CR at the time of DLI. In contrast, none of the five patients with aggressive disease (DLCL or Richter's transformation) achieved a CR despite receiving prior chemotherapy. Importantly, remissions induced by DLI therapy appear durable with only one patient relapsing at 18 months post-DLI with a median duration of follow-up of surviving patients of 40 months and with seven of the nine surviving patients in continuing CR >3 years post-DLI. Other studies of DLI have reported responses in lymphoma. Marks et al³ reported that out of 13 patients with follicular NHL, eight achieved CR post-DLI; however, the follow-up in these cases was short and no conclusion was made about the durability of these responses. Furthermore, they saw no responses in seven patients with CLL. Morris et al⁷ reported four out of seven responses in low grade NHL and in 1 of two patients with MCL. Dreger et al⁸ studied the effect of DLI in 12 patients with CLL and observed three complete and one partial responses.

In this study, the median number of CD3 cells to achieve CR for sibling transplants was 2 10⁷/kg whereas nonresponding patients received a higher number of cells at 5 10⁷/kg. This suggests that the lack of response in these patients was related to the underlying biology of the disease rather than to under treatment. The major toxicity of DLI therapy in our hands was acute GVHD that developed in 44% of patients. However, clinical responses were seen in the absence of acute GVHD in four out of 10 patients, although the majority of surviving patients have since developed mild chronic GVHD. Ways of improving the response of patients with aggressive disease are needed and may include the use of post-DLI chemo-immunotherapy to prevent disease progression in the interval before the development of a GVL effect. This strategy was successful in one patient with MCL who developed GVHD and achieved CR despite post-DLI dexamethasone and chemotherapy.

References

References