Multiple sclerosis flares associated with recombinant granulocyte colony-stimulating factor

Methods.

Recombinant human granulocyte colony-stimulating factor 10 to 16 μg/kg/day was administered subcutaneously or IV for 4 consecutive days. Apheresis commenced on day 5 and continued daily with rhG-CSF at the same dose until at least 6.5 × 10⁶ CD34⁺ cells/kg were collected. Peripheral blood stem cells were infused, usually 2 to 4 weeks later, after patients received high-dose immunosuppression therapy. Anti-thymocyte globulin 10 to 15 mg/kg was given IV for 3 to 6 days with methylprednisolone premedication just before and after stem cell infusion. rhG-CSF 5 μg/kg/day IV was restarted on day +1 after stem cell infusion and continued until the absolute neutrophil count was greater than 1.0 K/μL for 3 consecutive days.

Results.

Table 1 presents clinical details before protocol enrollment of the four patients whose MS flared during rhG-CSF therapy (Patients 1 through 4) and the six patients whose disease did not flare (Patients 5 through 10). MRI gadolinium enhancement was present in all three patients whose disease flared, compared with one of six patients whose disease did not flare. None of the patients had repeat MRI at the time of neurologic worsening. In Patients 1 through 4, there appeared to be neurologic worsening from a pre-existing lesion rather than development of new lesions (table 2).

Patient 1 became hemiplegic on the fifth day of rhG-CSF. Prior examination had shown upper motor neuron clumsiness and weakness, greater on the right side. Because of the neurologic worsening, the patient was taken off protocol and given methylprednisolone 1000 mg/day for 5 days. On the fourth day of methylprednisolone, there was improvement in right-sided strength, and in 3 more days her deficit had returned to the pre–rhG-CSF level.

Patient 2 was stable while receiving rhG-CSF for stem cell mobilization, during immunosuppression therapy with busulfan and cyclophosphamide, and while receiving rhG-CSF from day +1 to day +15. However, on day +25, her white cell count decreased to 1.7 K/μL and one dose of rhG-CSF at 5 μg/kg was given subcutaneously. The following day, the patient could no longer lift her right leg and could barely raise her right arm. With no improvement on day +27, she was admitted to the hospital and methylprednisolone was started at 1000 mg/day IV. Her deficit improved to the pretransplantation level after 3 days of methylprednisolone.

Patient 3 experienced symptomatic worsening of leg strength on the second day of rhG-CSF and became bed-bound 2 days later. Immunosuppression was given according to protocol with total-body irradiation and cyclophosphamide. Antithymocyte globulin with methylprednisolone premedication was then started, and six doses were administered 1, 3, and 5 days both before and after stem cell infusion. Leg strength and function began to improve 3 days after stem cell infusion. rhG-CSF was not given after transplantation in this patient.

Patient 4 had been ambulatory with a walker until 6 months before enrollment on the study, when she became paraplegic and experienced increasing weakness in her arms with a sensory level at C2-3. After completion of 6 days of rhG-CSF for stem cell collection, she experienced respiratory difficulties requiring intubation. There was no evidence of active infection or pulmonary embolism. Her clinicians attributed the breathing difficulty to the high cervical cord lesion and suspected that the combined effects of analgesics and progression of respiratory compromise after rhG-CSF resulted in pulmonary failure. Because of persistent respiratory instability, the patient was removed from the investigational study, and in accordance with her request, she was extubated. Death occurred 14 days after hospital admission for stem cell collection.

Discussion.

Neurologic problems associated with rhG-CSF mobilization in patients with MS have not, to our knowledge, been reported previously (Burt et al.1 and personal communication, Dr. Alan Tyndall of the EBMT and EULAR registry, 1999). Over 90% of patients with MS reported to the registry received cyclophosphamide just before rhG-CSF for stem cell mobilization. Our patients, however, and those of Burt et al.1 received rhG-CSF alone. The presence of gadolinium enhancement on the pretransplantation MRIs of patients whose disease later flared (see table 1) suggests that active MS may predispose to rhG-CSF–induced flares.

The possibility that these flares were fortuitous and not caused by rhG-CSF cannot be excluded. However, there have been reports in other autoimmune diseases of rhG-CSF triggering or worsening immune phenomena. These include anecdotal reports of rhG-CSF triggering cutaneous vasculitis and arthritis, as well as a report of exacerbation of neurologic symptoms in systemic lupus erythematosus.2 The mechanism is uncertain. In another patient with MS and in control subjects, we found that 5 days of rhG-CSF administration suppressed rather than activated peripheral blood lymphoproliferative responses to myelin protein epitopes (data not shown). Suppression may be mediated by monocytes, found in peripheral blood at twofold greater amounts after rhG-CSF, and reported to suppress alloantigen responses.3 rhG-CSF has also been shown to alter T cell subsets in favor of CD4⁻/CD8⁻ T cells,4 a population known in experimental systems to suppress graft-versus-host disease.

There is an extensive literature on cytokines in MS.5 In general, an increase in inflammatory cytokines—particularly tumor necrosis factor-α (TNFα), interferon-γ (IFNγ), and interleukin (IL)-12—correlates with disease activity, whereas an increase in anti-inflammatory cytokines—particularly IL-10—correlates with remission. In normal volunteers receiving rhG-CSF, ex vivo release of stimulated peripheral blood mononuclear cells showed increased IL-10 and decreased or attenuated TNFα and IFNγ.6 In addition, rhG-CSF treatment has been shown to shift splenocyte cytokines to a type 2 helper T-cell pattern in mice.7 However, another group measuring plasma levels reported a 25-fold increase in the type 1 helper T-cell cytokine TNFα in normal volunteers, with the rise starting within 24 hours of the first rhG-CSF dose.8 The role of cytokine release in MS disease activity in transplant recipients clearly needs further investigation.

Another possible mechanism involves a neutrophil effect. With daily rhG-CSF treatment, there is an eightfold increase in neutrophil count by 3 days, compared with only a 1.5- to 2-fold increase in T-cell counts. Neutrophils infiltrate the parenchyma in some patients after stem cell transplantation for oncologic disease, producing the so-called engraftment syndrome (fever, skin rash, capillary leak, and pulmonary infiltrates), a condition that occurs more often when rhG-CSF is used after transplantation.9 Corticosteroids also produce leukocytosis but, unlike rhG-CSF, corticosteroids decrease the capacity of neutrophils to migrate from the vascular space,10 and methylprednisolone is used therapeutically in the engraftment syndrome. Both the neurologic worsening in Patient 2 on day +25 (a time when the engraftment syndrome occurs) and this patient’s response to methylprednisolone support a possible role of neutrophil infiltration in regions of demyelinating plaques with local release of toxic mediators.

Three patients improved on methylprednisolone. Consequently, we have amended our protocols to include corticosteroid administration concurrently with rhG-CSF during stem cell mobilization. In addition, we believe it prudent to use a minimum duration of rhG-CSF to achieve early granulocyte recovery after transplantation and to avoid altogether post-transplantation rhG-CSF in those patients whose disease flared during mobilization.