Ultra-short course sirolimus contributes to effective GVHD prophylaxis after reduced-intensity allogeneic hematopoietic cell transplantation

Abstract

Reduced-intensity conditioning (RIC) allo-SCT is a potentially curative treatment approach for patients with relapsed Hodgkin’s or non-Hodgkin’s lymphoma. In the present study, 37 patients underwent RIC allo-SCT after induction treatment with EPOCH-F(R) using a novel form of dual-agent immunosuppression for GVHD prophylaxis with CsA and sirolimus. With a median follow-up of 28 months among survivors, the probability for OS at 3 and 5 years was 56%. Treatment-related mortality was 16% at day +100 and 30% after 1 year of transplant. Acute GVHD grades II–IV developed in 38% of patients, suggesting that the regimen consisting of CsA and an ultra-short course of sirolimus is effective in the prevention of acute GVHD.

INTRODUCTION

The use of reduced intensity conditioning (RIC) instead of myeloablative regimens has extended the indications for allo-SCT in patients with relapsed Hodgkin’s lymphoma (HL) or non-Hodgkin’s lymphoma (NHL). Older and medically less-suited patients, who would not be candidates for a standard allo-SCT, may now be offered a potentially curative treatment. The presumed advantage of allo-SCT is the graft-vs-lymphoma effect, which may explain why patients may achieve long-term disease-free survival (DFS). Previous reports have suggested that patients with various subtypes of NHL and HL may benefit from RIC allo-SCT. In general, the more indolent forms of NHL, like follicular lymphoma are better suited for this treatment.¹ However, promising results have been obtained with more aggressive diseases like mantle cell lymphoma,^2,3 T-cell lymphomas and diffuse large cell lymphoma.⁴ In HL, the evidence for a beneficial effect of RIC allo-SCT is less convincing because of higher relapse rate.⁵

Several reports have shown that lymphoma patients with chemo sensitive disease who are in a remission pre-transplant stand for a better chance of becoming long-term disease-free individuals.^6,7 Hence, an effective salvage regimen is likely to be important in order to achieve disease control. Furthermore, a significant depletion of host lymphocytes during induction promotes a more rapid full donor chimerism post transplant and thus potentially enhances the graft-vs-lymphoma effect. The dose-adjusted (DA) EPOCH-Fludarabin (F) regimen developed by Bishop et al.⁸ include fludarabine m in order to achieve this.

Transplant related mortality (TRM) remains a problem with allo-SCT, even with the use of RIC instead of myeloablative conditioning. GVHD, leading to organ toxicities and immunodeficiency, still represents an obstacle to a broader use of RIC allo-SCT. Single-agent calcineurin inhibitors as such CYA are commonly applied as GVHD prophylaxis. Recent studies have shown that prolonged administration of rapamycin (sirolimus) has remarkable clinical efficacy for GVHD prevention when used in combination with a calcineurin inhibitor.⁹ However, long-term sirolimus treatment is complicated by hyperlipidemia, pneumonitis and thrombotic microangiopathy.^10–12 To our knowledge, the duration of sirolimus therapy required for effective GVHD prophylaxis has not been established. Hence, in the present study, we aimed to determine whether the combination of CYA with an ultra-short course of sirolimus might represent effective acute GVHD prophylaxis in the RIC allogeneic hematopoietic cell transplantation setting.

We report the results of RIC allo-SCT in 37 patients after induction treatment with 1–3 cycles of DA EPOCH-F with addition of rituximab (R) for CD20-positive B-cell lymphomas and monitoring of tumor response and lymphodepletion before transplant. Dual-agent immunosuppression with CYA and an ultra-short course of sirolimus resulted in rapid donor chimerism and low rates of acute GVHD.

MATERIALS AND METHODS

Patients and donors

RIC allo-SCT in Norway is centralized to one institution, Oslo University Hospital. The current report includes 37 adult patients with relapsed HL or NHL who underwent RIC allo-SCT from 2005 to 2010. Peripheral blood stem cells were harvested from an HLA-identical family member or a 9/10 or 10/10 Ag-matched unrelated donor.

Patient characteristics at time of transplant are summarized in Table 1. All patients had relapsed disease after a median of 4 prior chemotherapy regimens (range 2–7), including auto-SCT in 27 cases. The median age was 53 years (range 21–67). Twenty-seven patients (73%) were male and 10 (27%) patients were female. Follicular lymphoma was the most common subgroup (10 patients), followed by HL (7 patients), diffuse large B-cell lymphoma (7 patients), transformed follicular lymphoma (6 patients), T-cell lymphoma (3 patients), mantle cell lymphoma (2 patients) and mycosis fungoides (2 patients). A matched family donor was used in 22 patients (60%) and a matched unrelated donor in 15 patients (40%). The source of stem cells was peripheral blood mobilized stem cells in all cases.

Table 1.

Patient characteristics

Characteristics	N (%) or median (range)
No. of patients	37
Patient age (years)	52 (19–67)
Patient sex
Male	27 (73)
Female	10 (27)
Disease
Follicular lymphoma	10
Hodgkins lymphoma	7
Diffuse large B-cell lymphoma	7
Transformed follicular Lymphoma	6
T-cell lymphoma	3
Mantle cell lymphoma	2
Mycosis fungoides	2
No line therapy before allo-SCT, median (range)	4 (2–7)
No. of patients with auto-SCT before allo-SCT	27 (73)
Donor type
HLA-matched family	22 (60)
HLA-matched unrelated	15 (40)
HCT co-morbidity index
0	14 (38)
1 or 2	17(46)
≥3	6(16)
Median follow-up for patients alive (range)	28 (14–78)

Abbreviation: HCT = hematopoietic cell transplant.

Induction treatment, conditioning and transplant

The majority of patients received induction therapy with 1–3 cycles of DA EPOCH-F,⁸ which consists of a 4-day continuous infusion of etoposide (50 mg/m²/day), vincristine (0.4 mg/m²/day) and doxorubicin (10 mg/m²/day); fludarabine (25 mg/m²/day), prednisolone (60 mg/m²/day) orally days 1–5 and CY (750 mg/m²) i.v. on day 5. Rituximab 375 mg/m² was given on day 2 in patients with CD20-positive NHL. PEG-fias given 6 mg was administered s.c. on day 6. Cycles were repeated every 21 days. The main purpose of adding fludarabine was to deplete host T cells and thus augment engraftment, secondarily to increase the efficacy of the regimen. Patients were evaluated with CT-scans, BM samples and counts of CD4-positive T cells in peripheral blood before next cycle. The aims were to reduce the CD4 counts below 0.100 × 10⁹/L before conditioning and to achieve the best possible remission. Prophylaxis with cotrimoxazole and fluconazole were given.

Patients received a conditioning regimen consisting of fludarabine 30 mg/m² and CY (600 mg/m²) on days −6 to −3. An initial cohort of seven patients that received a twofold higher dose of CY had a higher rate of acute pulmonary toxicity and as such, subsequent patients received the reduced dose. Allogeneic CD34+ cells with a minimum of 4 × 10⁶/kg recipient body weight were administered to all patients.

GVHD prophylaxis and grading

CYA was started on the day before transplant and was administered at therapeutic levels until day +100 and then tapered between day +100 and day +180 in the absence of GVHD. In addition, an ultra-short course of oral sirolimus was given, 12 mg on day −2 and then 4 mg daily for three more days. Serum levels of sirolimus were measured at day −1, day +2, day +6 and day +8. Acute GVHD was graded according to the criteria published from the consensus conference on acute GVHD grading.¹³ Chronic GVHD was designated as limited or extensive.¹⁴

Co-morbidities

Patient co-morbidities were assessed pre-transplant by review of the medical records and using the definitions of the 17 co-morbidities included in the hematopoietic cell transplantation co-morbidity index described by Sorror.^15,16

Statistics

Survival and time of relapse were calculated from the date of transplantation until date of death, relapse or last follow-up, as appropriate, through July 2011. Statistics were performed using the SPSS 16 software package (SPSS Inc., Chicago, IL, USA). Probabilities of OS and DFS were calculated with the method of Kaplan–Meier. The curves for DFS were calculated using the definitions of current leukemia-free survival, because of the possibility of achieving durable remissions after relapse.¹⁷

RESULTS

Induction treatment with DA EPOCH-F(R)

Thirty-six patients received pre-transplant treatment with DA EPOCH-F(R) over 1–3 courses, which resulted in significant lymphodepletion in all patients. Median absolute lymphocyte count decreased from 0.7 × 10⁹/L (0.1–4.8) to 0.2 × 10⁹/L (0.01–0.6) and median CD4 count decreased from 0.24 × 10⁹/L (0.04–1.74) to 0.1 × 10⁹/L (range 0.01–0.31). EPOCH-F(R) proved to be a very effective regimen in this heavily pretreated group of patients. Seventy-eight percent of the patients responded to induction treatment (CR 30%, PR 48%) and the remaining had stable disease. The response rates following EPOCH-F(R) and subsequent allo-SCT are summarized in Table 2. This regimen was well tolerated with clinically manageable hematological toxicities; grade III–IV neutropenia occurred in 28 patients and grade III–IV thrombocytopenia was observed in 30 patients. Fourteen patients developed infectious complication grade III–IV during treatment with EPOCH-F(R); all were successfully treated.

Table 2.

Median T-cell counts before and after EPOCH-F(R)

	Before EPOCH-F(R)	After EPOCH-F(R)
CD3+ T cells	0.7 × 109/L (range 0.1–4.8)	0.2 × 109/L (range 0.01–0.6)
CD4+ T cells	0.24 × 109/L (range 0.04–1.74)	0.01 × 109/L (range 0.01–0.31)
CD8+ T cells	0.41 × 109/L (range 0.04–3.04)	0.1 × 109/L (range 0.01–0.39)

Sirolimus pharmacokinetics and side effects

Serum sirolimus levels were monitored on days −1, day +2, day +6 and day +8. Median sirolimus levels were 5.3 ng/mL (day −1; range 4.1–7.4; n = 9), 5.4 ng/mL (day +2; range 7.0–4.3; n = 8) and<2.5 ng/mL (as measured on days +6 and +8, n = 8).

There was one possible case of transplant-associated microangiopathy (TAM), but no cases observed of pneumonitis, sinusoidal obstruction syndrome.

Engraftment and donor chimerism

There were no cases of non-engraftment. The median donor chimerism in unseparated cells at day 30 was 95% (range 22–99) and was 98% in purified T cells (60–99). Median donor chimerism at day 100 was >99% (range 47 to >99) in both unseparated cells and in T cells (range 90 to >99).

TRM and organ toxicities

Although RIC extends the availability of allo-SCT to a larger group of patients, it is still associated with significant morbidity and mortality. TRM owing to GVHD, infection or toxicity was seen in 10 patients (27%) (Table 3). TRM rate was 16% at day 100 and 30% at 1 year. Of the initial cohort of seven patients who received the higher dose of CY (1200 mg/m²/day), three experienced serious pulmonary toxicity, which was fatal in two cases. Another patient died of pulmonary complications with signs of bronchiolitis obliterans syndrome 1 year after transplant. Subsequently, the dose of CY was reduced to 600 mg/m²/day, resulting in better tolerability of the regimen with no acute pulmonary toxicity and lower TRM rate. Of n = 30 patients treated with the lower dose, there were two cardiac related deaths (days +24 and +76), three cases of multiorgan failure (days +67 and +194), one death caused by infection (days +231) and one case of end-stage pulmonary disease as a consequence of bronchiolitis obliterans syndrome (day +915). In this cohort, the TRM rates at day +100 and 1 year were 10% and 23%, respectively. Four out of six patients who died before evaluation at day +100 underwent autopsy and none showed evidence of lymphoma. One patient developed kidney failure requiring dialysis and potential kidney transplantation.

Table 3.

Summary of treatment responses and outcomes

Patient no.	Diagnosis	Donor	EPOCH-F(R) response	Response after tx	Acute GvHD (grade)	Chronic GVHD (grade)	Relapse (months)	Outcome (months of FU)	Cause of TRM
30	FL	MUD	SD	CR	II	0	N/A	Alive in CR (22)
2	FL	Family	PR	CR	0	Limited	N/A	Alive in CR (78)
6	FL	Family	PR	CR	0	N/A	N/A	Died in CR (2)	MOF
7	FL	Family	CR	CR	0	0	N/A	Alive in CR (70)
10	FL	Family	PR	CR	0	0	N/A	Alive in CR (43)
11	FL	Family	SD	PR	II	Limited	Relapse (22)	Died in relapse (36)
18	FL	MUD	PR	CR	II	Extensive	N/A	Alive in CR (28)
33	FL	Family	SD	N/A	IV	N/A	N/A	Died (2)	MOF
35	FL	MUD	PR	CR	0	Limited	N/A	Alive in CR (15)
36	FL	MUD	PR	CR	0	Extensive	N/A	Alive in CR (14)
3	tFL	Family	PR	CR	III	N/A	N/A	Died in CR (3)	MOF
4	tFL	Family	CR	CR	0	0	N/A	Alive in CR (75)
5	tFL	Family	PR	CR	0	Limited	N/A	Alive in CR (75)
23	tFL	Family	CR	CR	0	0	N/A	Alive in CR (26)
32	tFL	Family	PR	CR	0	0	N/A	Alive in CR (21)
37	tFL	MUD	PR	CR	II	Limited	N/A	Died in CR (8)	Infection
25	DLBCL	MUD	SD	CR	0	N/A	N/A	Died in CR (1)	Cardiac
28	DLBCL	MUD	CR	CR	II	Limited	N/A	Alive in CR (24)
20	DLBCL	Family	N/A	CR	0	Limited	N/A	Alive in CR (28)
13	DLBCL	Family	SD	SD	III	Extensive	Relapse (4)	Died in relapse (10)
14	DLBCL	Family	PR	SD	II	Extensive	Relapse (3)	Died in relapse (9)
16	DLBCL	MUD	N/A	CR	0	Extensive	N/A	Alive in CR (33)
17	DLBCL	Family	SD	CR	III	N/A	N/A	Died in CR (3)	MOF
26	HL	MUD	CR	CR	III	Extensive	N/A	Died in CR (6)	MOF
27	HL	MUD	CR	CR	0	Limited	Relapse (23)	Alive in relapse (24)
15	HL	MUD	CR	PR	III	Extensive	N/A	Died in CR (31)	GVHD
34	HL	MUD	SD	PR	0	Limited	Relapse (7)	Alive in relapse (17)
21	HL	MUD	SD	PR	III	Limited	Relapse (4)	Alive in relapse (28)
8	HL	Family	CR	CR	0	Limited	N/A	Alive in CR (61)
9	HL	Family	SD	CR	0	Extensive	Relapse (14)	Alive in CR (46)
31	TCL	Family	PR	CR	1	Limited	N/A	Alive in CR(21)
19	TCL	Family	PR	CR	0	0	N/A	Alive in CR (28)
24	TCL	Family	PR	CR	0	0	N/A	Alive in CR (26)
1	MF	Family	PR	CR	0	Extensive	N/A	Died in CR (12)	GVHD
29	MF	MUD	PR	N/A	0	N/A	N/A	Died (3)	Cardiac
22	MCL	MUD	PR	CR	II	Limited	N/A	Alive in CR (27)
12	MCL	Family	CR	CR	0	Limited	Relapse (23)	Alive in CR (41)

Abbreviations: DLBCL = diffuse large B-cell lymphoma; FL = follicular lymphoma; FU = follow-up; HL = Hodgkin’s lymphoma; MCL = mantle celle lymphoma; MF = mycosis fungoides; MOF = multi-organ failure; MUD = matched unrelated donor; ND = not determined; N/A = not applicable; ORR = overall response rate; SD = stable disease; tFL = transformed follicular lymphoma; TRM = transplant-related mortality. Bold patient numbers indicate the higher dose of CY.

Nine patients (24%) developed a post-transplant CMV reactivation with a median time to reactivation of 52 days (range 21–195). Four patients developed proven or probable invasive fungal infections, which contributed to death in two of them. Significantly, 14 out of 27 (48%) survivors have developed hypogammaglobulinemia requiring substitution.

Acute and chronic GVHD

Acute GVHD grades II–IV was observed in 14 patients of the 37 evaluable patients (38%) and grade III–IV in 7 patients (19%). This is significantly less than that observed in a cohort of lymphoma patients (68%) reported earlier and treated with identical pre-transplant chemotherapy, a conditioning regimen containing fludarabine and CY, and single-agent CYA¹⁸ (P = 0.01). Chronic GVHD was observed in 22 out of 31 (71%) patients surviving past day 100; 9 of these experienced extensive chronic GVHD, including 2 patients with bronchiolitis obliterans syndrome. At last follow-up, 12 out of 24 evaluable patients were still on immunosuppressive treatment, 11 with active GVHD.

The co-morbidity index was assessed pre-transplant. Fourteen patients had a hematopoietic cell transplantation-specific comorbidity index (HCT-CI) score of 0, 17 patients had a score of 1–2 and 6 patients had a score of ≥3. There was a trend towards a poorer outcome in patients with significant co-morbidities. Eleven out of 14 patients with a HCI-CI score of 0 are alive at last follow-up, compared with, 9/17 patients and 4/6 patients in the HCT-CI groups 1–2 and >3, respectively.

Outcome after transplant

By July 2011, with a median follow-up of 28 months, 24 patients (65%) were alive and 21 of these were in CR. The probability of OS and DFS at 3 and 5 years was 56% (Figure 1). A total of seven patients developed histologically verified relapses and one patient developed clinical signs of relapse with PET-positive lesions in the mediastinum. Five of the relapses were in the family donor group, whereas three were in the matched unrelated donor group. Two of the patients who relapsed responded to therapeutic interventions and were brought back in long-term CR. One patient with mantle cell lymphoma had a relapse in a supraclavicular lymph node and the left tonsil 9 months after transplant, which was treated with tonsillectomy, radiation therapy and a reduction in the immunosuppressive treatment. He is in CR 3 years following allo-SCT. The other patient presented with PET-positive lympha-denopathy in the right hilar region giving suspicion of relapse 1 year after transplant for Hodgkin’s disease. Biopsy attempts were unsuccessful. Immunosuppression was reduced with subsequent complete regression of the PET-positive lesion. He is in CR but with extensive chronic GVHD 3 years after transplant. These two patients were subsequently reclassified as disease free and are not represented as relapsed in curves for DFS. Except for one relapse among patients with follicular lymphoma, all other cases where local treatment and/or tapering of immunosuppression did not bring the patients back to stable remission occurred among patients with diffuse large B-cell lymphoma (two cases) and HL (three cases). During follow-up, there were three deaths due to relapse.

Figure 1.

OS (a) and DFS (b) after RIC allo-SCT for relapsed malignant lymphoma.

DISCUSSION

Allo-SCT with RIC is a potentially curative treatment modality for patients with relapsed malignant lymphoma where other options do not have realistic potential for long-term survival. Use of RIC protocols have extended the option of allo-SCT to patients who were previously excluded because of unacceptable risks of TRM. In the present study we aimed to induce the best possible remission and a significant depletion of host lymphocytes in order to provide disease control and promote rapid donor chimerism after transplant. This strategy would be expected to be associated with lower rates of relapse after RIC allo-SCT.^8,19,20 By using the DA-EPOCH-F(R) induction regimen we achieved high response rates of 78% in these heavily pretreated lymphoma patients, of which the majority had relapsed after auto-SCT. Moreover, the toxicity profile of this regimen was manageable. DA-EPOCH-F(R) combined with the conditioning resulted in lymphodepletion that was comparable to what could be expected after myeloablative regimens, but with less toxicity. In accordance with this all patients engrafted and the majority of patients achieved complete and durable donor chimerism at day +30 after transplant. The results are in line with previous reports in malignant lymphoma and multiple myeloma.^18,19

Significant improvement in the quality of response was seen after transplant, the majority of patients achieving durable CR. We have only observed eight relapses so far and two of these were brought back to stable remissions by tapering of immunosuppression without systemic lymphoma therapy, most likely reflecting a GVL response. A third patient was brought into CR after chemotherapy and tapering of immunosuppression. There is clear evidence of a GVL effect in malignant lymphoma,^14,21 however, not as prominent as reported in chronic myelogenous leukemia.²²

Probability for overall and DFS at 3 and 5 years was 56% in our study, which included patients with different subtypes of NHL and HL. This compares favorably with earlier reports on RIC allo-SCT in malignant lymphoma.^23–26 In line with previous studies, more relapses were observed among patients with diffuse large B Cell lymphoma and HL. Moreover, late relapses were more common in the HL group. Interestingly, we saw no significant difference in overall or DFS with the use of alternative donors compared with a family donor.

The primary objective of this study was to reduce rates of acute GVHD by using a novel dual-agent immunosuppressive regimen containing CsA and an ultra-short 4-day course of sirolimus. The rates of grade II–IV acute GVHD was 37%, which appears to be favorable relative to the 68% rate observed by Dean et al.¹⁸ using identical pre-transplant chemotherapy, a conditioning regimen containing fludarabine ide and CY, and single-agent CYA as GVHD prophylaxis. Although the rate of acute GVHD reported by Dean et al.¹⁸ is somewhat higher than what is commonly reported, we conclude that the addition of ultra-short course of sirolimus may have an important role in acute GVHD prophylaxis. It is possible that the lower dose of CY conditioning in our study may have also contributed to a relatively favorable acute GVHD rate. However, the fact that only one out of six evaluable patients (17%) in our initial cohort of patients who received the higher dose of CY developed acute GVHD (grade III) makes this less likely. Sirolimus is well known as an effective agent for treatment and prevention of GVHD. It is most commonly used concomitantly with tacrolimus, which is known to act synergistically with sirolimus in GVHD prevention.²⁷ Sirolimus has several theoretical advantages in prevention of GVHD, including inhibition of Ag presentation and DC maturation and may contribute to upregulation of T-regulatory cells²⁸; our study did not address the potential mechanism of GVHD prevention afforded by ultra-short course sirolimus. Toxicity has been an issue, especially endothelial cell-related complications such as sinusoidal obstruction syndrome following BU-based conditioning regimens and TAM. A TAM incidence of 10.8% was reported by Cutler et al.¹¹ in sirolimus-containing immunosuppressive regimens. As such, the long-term use of sirolimus in GVHD prophylaxis may carry significant risks for toxicity and it was therefore given in a short course of 4 days in the peritransplant period and discontinued after day +1 post transplant; using this regimen, we observed therapeutic sirolimus levels only through week 1 of transplant. There was one probable case of TAM and no registered cases of sinusoidal obstruction syndrome in our patient group, indicating that this combination of immunosuppressive drugs may not contribute to endothelial cell-related side effects. Chronic GVHD remains a significant source of morbidity with rates of 71% in the present study with and 29% developing extensive cGVHD.

Using the higher dose of CY of 1200 mg/m²/day in the first seven patients, we encountered problems with excessive toxicity, especially pulmonary toxicity. This group of patients was heavily pretreated with multiple courses of chemotherapy and auto-SCT. Some had also received radiotherapy. As such, it is likely that these patients would be more prone to the toxic effects of new chemotherapy. In a previous report where the same conditioning regimen was used 37% developed acute pulmonary toxicities.²⁹ After dose reduction to 600 mg/m²/day was applied for the remaining 30 patients we did not observe major acute noninfectious pulmonary problems. For this cohort the TRM rate at day +100 and 1 year was acceptable at 10 and 23%, respectively. Hence, the data suggested that the lower dose of 600 mg/m² CY was more tolerable with regard to immediate toxicity without compromising engraftment.

Almost half of the patients developed hypogammaglobulinemia dependent on substitution during follow-up after transplant. This is a higher frequency than commonly reported after allogeneic transplants.^30,31 One reason for this could be the nature of the underlying lymphoid malignancy. Another contributing factor could be that the majority of patients in our report had received treatment with the B-cell depleting monoclonal anti-CD20 Ab rituximab as single-agent or in combination with chemotherapy. Some publications have linked hypogammaglobulinemia to the presence of chronic GVHD.³⁰ Nine out of 13 patients with active chronic GVHD required gammaglobuline substitution, compared with only 4 out of 14 patients without active chronic GVHD.

In conclusion, we report that the DA EPOCH-F(R) induction regimen followed by RIC allo-SCT results in significant lymphodepletion and high response rates before transplant and rapid engraftment, complete donor chimerism and durable responses after transplant. GVHD prophylaxis with CYA in combination with an ultra-short course of sirolimus resulted in lower rates of acute GVHD than previously reported with CYA alone. Future studies evaluating this combination should consider extending the period of sirolimus treatment in order to further reduce the rate of acute GVHD with acceptable toxicities.