Predictors of response to rituximab in patients with neuropathy and anti ^ myelin associated glycoprotein immunoglobulin M

Transcription

1 Journal of the Peripheral Nervous System 12: (2007) RESEARCH REPORT Predictors of response to rituximab in patients with neuropathy and anti ^ myelin associated glycoprotein immunoglobulin M Luana Benedetti 1, Chiara Briani 2,MarinaGrandis 1,Tiziana Vigo 1, Marco Gobbi 3, Elisabetta Ghiglione 1, Marinella Carpo 4, Dario Cocito 5, Christina M. Caporale 6,MariaP.Sormani 7, Giovanni L. Mancardi 1, Eduardo Nobile-Orazio 8, and Angelo Schenone 1 1 Department of Neurosciences, Ophthalmology and Genetics, University of Genova, Genova; 2 Department of Neurosciences, University of Padova, Padova; 3 Department of Internal Medicine and Medical Specialties, University of Genova, Genova; 4 Department of Neurological Sciences, Fondazione IRCCS, Ospedale Maggiore Policlinico, Mangiagalli Regina Elena, Milano; 5 Neurologia I, Department of Neurosciences, A.S.O. San Giovanni Battista, Torino; 6 Neuromuscular Diseases Unit, Ce.S.I, Foundation University G. D Annunzio, Chieti; 7 DISSAL, Unit of Biostatistics, University of Genova, Genova; 8 Department Neurological Sciences, Milan University, IRCCS Humanitas Clinical Institute, Rozzano, Milano, Italy Abstract We evaluated the efficacy and safety of rituximab in an open-label, uncontrolled study of 13 patients with polyneuropathy associated with antibodies to myelinassociated glycoprotein (MAG) and correlated the response to therapy with clinical and laboratory features. One year after rituximab therapy, anti-mag immunoglobulin M (IgM) titers were significantly reduced. At that time, eight patients (62%) had improved in both the inflammatory neuropathy cause and treatment (INCAT) sensory sumscore and the Medical Research Council sumscore for muscle strength and seven of them also in the INCAT disability score. The improvement in the mean INCAT sensory sumscore was significant at 12 months and correlated with lower anti-mag antibody at entry and at followup. This study suggests that rituximab may be efficacious in patients with anti MAG associated neuropathy and particularly on sensory impairment and in those with moderately elevated antibody titers. These findings suggest that antibody reduction below a critical level may be necessary to achieve clinical improvement. Key words: anti-mag antibodies, anti-mag polyneuropathy, IgM, monoclonal gammopathy, rituximab Introduction Polyneuropathy associated with anti myelin associated glycoprotein (MAG) antibodies is the best defined and most frequent immunoglobulin M (IgM) paraproteinaemic Address correspondence to: Luana Benedetti, MD, Department of Neurosciences, Ophthalmology and Genetics, University of Genova, Via De Toni 5, Genova, Italy. Tel: þ ; Fax: þ ; luanabenedetti@libero.it neuropathy (Nobile-Orazio, 1998). Several lines of evidence suggest that anti-mag IgM causes the neuropathy through a complement-mediated mechanism (Monaco et al., 1990). This has prompted the use of immune therapies, based on the assumption that reducing antibody levels will improve the neuropathy. A recent Cochrane review concluded, however, that there is inadequate evidence in anti- MAG paraproteinaemic neuropathy to recommend any treatment (Lunn and Nobile-Orazio, 2003). ª 2007 Peripheral Nerve Society 102 Blackwell Publishing

2 Rituximab is a chimeric monoclonal antibody that specifically binds CD20 antigen on B-cells. This surface membrane marker is present from the pre-b cell development phase until differentiation to the mature B-cell. Rituximab induces antibody-dependent cellmediated and complement-mediated cytotoxicity of CD20þ B-cells (Maloney et al., 2002). A beneficial effect of rituximab on neuropathy associated with anti-mag antibodies has been reported in two openlabel, uncontrolled studies (Pestronk et al., 2003; Renaud et al., 2003; 2006) and in an abstract from a placebo-controlled study (Dalakas et al., 2006). In an attempt to identify possible predictors of the response to this therapy, we analyzed the effect of rituximab in 13 patients with anti-mag polyneuropathy consecutively treated in six Italian centers. Patients and Methods The present study was an uncontrolled open-label trial conducted by members of the Italian Peripheral Nerve Study Group. The study population included 13 patients affected by anti-mag polyneuropathy. They were highly variable in anti-mag antibody titers, grade of disability at baseline, disease duration, and associated hematological disease (Table 1). In all patients, routine laboratory tests including blood glucose, serum thyroid hormones, vitamin B 12, tumor markers, and hepatitis C serology, excluded other possible causes of neuropathy. Motor nerve conduction velocity (MCV), compound motor action potential (CMAP) amplitude, and distal motor latency (DML) of the ulnar and peroneal nerves were recorded. Orthodromic sensory nerve conduction velocity and sensory action potential amplitude of the sural nerve were measured. All patients had symmetrically prolonged distal latencies and reduced nerve conduction velocities (Table 2). Five patients had no previous treatment for the neuropathy; eight were unresponsive to other therapies. None was treated in the 3 months prior to study entry. Patients were enrolled after giving written informed consent. Rituximab was administered at a dosage of 375 mg/m 2 as intravenous weekly infusion for four consecutive weeks. Two patients received a second cycle of rituximab after 9 12 months. All patients underwent periodic clinical, immunological, and electrophysiological evaluations during the follow-up of 12 months. The neurological assessment was performed in each center by the same investigator at baseline and 1, 3, 6, and 12 months after the last treatment infusion, using the Medical Research Council (MRC) sumscore (Kleyweg et al., 1991) for muscle strength, the inflammatory neuropathy cause and treatment (INCAT) scales for arm and leg disability (Hughes et al., 2001), and the INCAT sensory sumscore (ISS) (Merkies et al., 2000). We considered a responder a patient who improved by one point in at least two scales. Nerve conduction studies were performed at baseline and 6 and 12 months after treatment. The neurophysiological examinations were performed in each center by the same operator according to normal data corrected for age, respecting the same distances. Foot and hand skin temperature was controlled at C. A change 10% in nerve conduction velocity was considered to be clinically Table 1. Clinical and hematological parameters of the patients at baseline and 12 months after therapy.* Patient, sex, age Neuropathyy duration (years) Hematological disease IgM (mg/dl)z Anti-MAGz ISSz MRCz INCATz 1, M, Thrombocytosis 200^100 1/100000^1/12800x 7^3 58^60 3^0 2, M, 61 5 MGUS IgM 600^580 1/800000^1/25600x 12^12 54^54 4^4 3, F, Waldenstrom 956^579 1/ ^1/800000x 10^11 52^48 6^6 4, F, 60 4 MGUS IgM 300^96 1/51200^1/51200x 11^10 58^59 3^3 5, F, 69 1 MGUS IgM 384^175 1/51200^1/1000x 8^2 58^60 2^1 6, F, 53 2 MGUS IgM 305^179 1/3200^1/1000x 9^6 59^60 3^2 7, M, 65 3 MGUS IgM 887^394 1/1600^0x 10^5 59^60 2^1 8, M, 55 4 Waldenstrom 1347^995 1/ ^1/51200x 10^12 58^60 2^2 9, M, 62 2 MGUS IgM 788^550 1/51200^1/3200x 9^8 46^50 8^7 10, M, Waldenstrom 827^685 1/25600^1/51200x 5^5 56^52 2^5 11, M, 70 4 Waldenstrom 1300^ ^35000k 9^6 54^60 3^1 12, M, 67 1 MGUS IgM 600^ ^70000k 8^10 56^54 2^4 13, F, 68 3 Waldenstrom 1900^ ^12500k 10^6 48^52 4^3 IgM, immunoglobulin M; MAG, myelin-associated glycoprotein; MGUS, monoclonal gammopathy of undetermined significance; MRC, Medical Research Council sumscore; INCAT, theinflammatory neuropathycause and treatment arm andlegdisability scores; ISS, INCATsensory sumscore; M, male; F, female. *The parametersinitalics correspondtoimprovement. yfromthe onsetofneurologicalsymptoms. zthe firstnumberrefers tovalues atbaseline, andthe secondnumberrefers tovalues at12 months. xwesternblot. kenzyme-linkedimmunosorbant assay. 103

3 Table 2. Electrophysiological findings at baseline and 12 months after therapy.*y Peroneal nerve Ulnar nerve MCV(m/s) MCV(%) DML(ms) DML(%) CMAP(mV) CMAP(%) MCV(m/s) MCVx (%) DML (ms) DML (%) CMAP (mv) CMAP (%) Patient ^21.0 þ ^26.6 þ ^0.3 þ ^33.2 þ10 7.2^5.9 þ18 7.2^7.5 þ4 2 NE NE NE NE NE NE 13.1^22.9 þ75 9.1^8.0 þ12 2.3^ NE NE NE NE NE NE 12.8^ ^ ^3.5 þ ^8.7 þ7 25.0^24.2 þ ^ ^34.7 þ22 9.4^ ^6.0 þ ^19.0 þ7 11.8^7.5 þ ^1.4 þ ^60.0 þ1 4.0^3.8 þ5 9.8^11.5 þ ^33.5 þ ^10.4 þ ^ ^ ^3.5 þ40 8.6^ NE NE NE NE NE NE 34.1^57.5 þ68 4.5^4.4 þ2 4.3^8.4 þ ^23.0 þ ^11.4 þ ^ ^ ^4.5 þ22 4.7^5.4 þ toNE toNE to NE ^ ^ ^ to NE to NE to NE ^29.1 þ2 5.7^ ^ NE NE NE NE NE NE 12.0^25.0 þ ^ ^ ^ ^11.8 þ ^ ^ ^4.0 þ2 4.0^ NE NE NE NE NE NE 24.0^33.5 þ ^12.8 þ5 2^2.8 þ40 MCV, motornerve conductionvelocity; DML, distalmotorlatency; CMAP, compoundmotoractionpotential; NE, notelicitable. *Theparametersinitalics correspondto animprovement 10 %. yforeachparameter, the firstcolumnreports thevalues atbaseline andat12 months, whereas the secondcolumnrefers tothepercentage ofimprovementcomparedwiththebaseline. relevant. Two patients were also clinically and electrophysiologically assessed after 24 months. Hematological and immunological tests were repeated at each visit and included: complete blood count, peripheral blood immunophenotype by flow cytometry, CD19þ B-cells, serum protein electrophoresis, and IgM levels. Anti-MAG antibodies titers were determined at the same time by Western blot (Nobile-Orazio et al., 1994) in 10 patients and by enzyme-linked immunosorbant assay (Buhlmann Laboratories AG) in three patients. In each patient, antibody titers were measured using the same method throughout the follow-up. Renal and liver functions were monitored. Side effects were recorded. Statistical analysis Statistical analysis was performed after 12 months in all the patients. The comparison between follow-up and baseline values were analyzed using the nonparametric Wilcoxon test for paired data and correlations evaluated by the nonparametric Spearman rank correlation coefficient. Results Immunological evaluation Circulating CD19þ B-cells were serially measured in seven patients, and all were undetectable after 1 month, reappeared at 6 months, and returned to pretreatment values after 1 year (data not shown). Median serum IgM levels decreased by 39% (p, 0.05) after 12 months, whereas anti-mag antibody titers were reduced by 68% after 1 month (data not shown) and by 87% at 12 months (p, 0.05) (Table 1). In the two patients with longer follow-up (patients 1 and 2), antibody titers at 24 months had a twofold increase compared with month 12 but were still lower than at baseline. Clinical evaluation Twelve months after therapy, eight patients (62%) were improved by at least one point in the ISS, two (15%) were stable and three (23%) had worsened, two within 1 month of treatment, and one after 6 months of stabilization. A mean overall improvement of 1.9 point (20%) in the ISS was observed at 12 months compared with baseline (p ¼ 0.05) (Table 1). The MRC sumscore at 12 months improved by at least one point in nine (69%), remained unchanged in one (8%), and worsened in three (23%) patients. The INCAT disability score improved by at least one point in seven (54%), remained unchanged in four (31%), and worsened in two (15%) patients. The mean improvement in both scales, however, was not significant. 104

4 Overall, eight patients (62%) had improved in at least two scales (patients 1, 4 7, 9, 11, and 13) including, in seven of them, the INCAT disability scale. Patients 1, 5 7, 9, 11, and 13 improved in ambulation because of a reduction in sensory ataxia and increase in muscle strength in the lower limbs. Patients 9, 11, and 13 also improved in other daily activities, such as handling knife and fork, washing hair, doing and undoing zippers, consistent with the improvement in MRC and ISS in the upper limbs. Patient 4 had a slight improvement in MRC and ISS, but this did not correspond to a reduction of the overall disability. Three patients had worsened (23%) (patients 3, 10, and 12) and two were unchanged (15%) (patients 2 and 8). Tremor in the upper limbs was reduced after therapy in two of the five affected patients. Of the two patients followed for up to 2 years, one (patient 1) maintained the improvement until 24 months, whereas the other (patient 2) remained stable for 18 months and then started to worsen (data not shown). Electrophysiological studies Analysis was performed on the most affected side. Sural nerve conduction velocities were not recordable in 12 patients. The peroneal CMAP could not be obtained in 5 of the 13 patients. Therefore, the evaluation was carried out on eight peroneal nerves and on all ulnar nerves and the results compared with baseline values (Table 2). In three patients, the peroneal nerve MCV improved 10%, whereas in two, it worsened and in three, it remained stable (range 0 9%). The peroneal nerve distal CMAP amplitude slightly increased in only two patients. The ulnar nerve MCV improved 10% in six patients; in three, it worsened; and in four, it was unchanged. The ulnar nerve distal CMAP amplitude also improved in six patients. The DML of peroneal and ulnar nerve improved 10% in four patients. Correlations Based on the clinical results showing a significant improvement at 12 months in the ISS only, we correlated ISS improvement with disease duration, IgM level, anti-mag antibody titer, and ulnar nerve MCV and distal CMAP amplitude (as possible expression of axonal damage), before therapy. Improvement of the ISS at 12 months significantly correlated with lower anti-mag antibody titers at baseline (R ¼ 0.67, p ¼ 0.03) (Fig. 1). None of the eight patients who improved after therapy had anti- MAG IgM titers at baseline higher than 1/100,000 (median 1/51,200) as compared with three of the five nonresponding patients (median 1/800,000). ISS change We also found a high correlation coefficient between ISS improvement and shorter disease duration (R ¼ 0.41, p ¼ 0.16), lower IgM level (R ¼ 0.63, p ¼ 0.08), higher ulnar MCV (R ¼ 0.5, p ¼ 0.2), and CMAP amplitude (R ¼ 0.66, p ¼ 0.07) at baseline, but these correlations were not significant. Improvement in ISS after therapy did not significantly correlate with changes in any of the electrophysiological parameters examined even if there was a positive, though not significant, correlation between ISS improvement, and ulnar MCV increase at 12 months (R ¼ 0.36, p ¼ 0.22). There was no correlation between the improvement of the ISS and the hematological diagnosis (monoclonal gammopathy of undetermined significance vs. Waldenstrom s macroglobulinemia) (p ¼ 0.84). There was also no significant difference between the mean severity of neuropathy at entry expressed as ISS in the responders (9.1) and nonresponders (9). Side effects No major adverse effects were registered. During the first infusion, one patient complained of headache that disappeared with a slower infusion rate, and another patient had chills, which responded to steroid therapy. One patient developed a transient mild leukopenia at 6 months. Discussion Ab anti-mag baseline (log scale) Figure 1. Correlation between response to therapy and anti myelin associated glycoprotein immunoglobulin M antibody titers at baseline. Improvement in INCAT sensory sumscore at 12 months significantly correlated with lower antibody titers at baseline (R ¼ 0.67, p ¼ 0.03). Only patients whose anti-mag titer was evaluated by Western blot were included in the analysis. Recently, rituximab has been suggested as a treatment for anti-mag antibody-associated polyneuropathy

5 (Pestronk et al., 2003; Renaud et al., 2003; Dalakas et al., 2006) on the assumption that B-cell depletion may result in a decrease in antibody production. This neuropathy is theoretically the ideal target of an anti- CD20 monoclonal antibody because it is deemed to be strictly related to an antibody-mediated mechanism (Monaco et al., 1990). Moreover, no consistently effective therapy is so far available for anti-mag polyneuropathy (Lunn and Nobile-Orazio, 2003). In this open-label study, 8 of the 13 patients (62%) with anti-mag polyneuropathy treated with rituximab had improved 1 year after treatment in both sensory (ISS) and motor (MRC) impairment and seven of them also in disability (INCAT). A similar frequency of response to rituximab, 66 and 75%, respectively, was also reported by Renaud et al. (2003) in their series of nine patients and by Dalakas et al. (2006) in their preliminary results of a placebo-controlled trial. In our study, however, only the mean ISS showed a significant improvement, whereas the changes in the mean MRC sumscore and INCAT disability score were not significant. This may be related to the relatively small number of treated patients and to the fact that because patients with anti-mag neuropathy mainly have a sensory impairment, this is probably more prone to reflect a response to therapy. An improvement in sensory impairment was also reported by Renaud et al. (2003). Other authors also demonstrated a significant improvement (22%) in strength after Rituximab treatment (Pestronk et al., 2003), possibly because they used a more sensitive method (hand dynamometry) to measure strength. Clinical improvement was associated in all but one of our patients with a consistent decrease in anti- MAG IgM titers. This was, however, also observed in three nonresponding patients who had much higher baseline titers. As also previously reported (Renaud et al., 2003), clinical improvement did not always correlate with electrophysiological improvement and vice versa, as only four of the eight patients who clinically improved had a consistent improvement in the electrophysiological parameters examined, whereas three had variable response and one consistently worsened. Electrophysiological improvement was usually more evident in the ulnar nerve than in the peroneal nerves, possibly reflecting the much higher degree of impairment in the latter nerve, which was often inexcitable. In an attempt to identify predictors of response to therapy, which might help with planning future therapeutic strategies, we correlated the improvement in the only score significantly changed after therapy (ISS) with a number of baseline parameters. The presence of low anti-mag antibody titers at baseline was the only parameter that was significantly associated with ISS improvement. This may explain why Broglio and Lauria (2005) reported a patient with extremely high titers of anti-mag antibodies who did not respond to rituximab. It is possible to speculate that low titers are more easily reduced below the critical level over which they are pathogenic. This hypothesis is also supported by the fact that in all but one of our patients responding to therapy, anti-mag titers were reduced after therapy to less than 1/51,200. Similarly, the median anti-mag titer 1 year after therapy in responsive patients was 1/6,400 compared to 1/51,200 in nonresponsive patients. Even if none of the other factors examined, such as disease duration and severity, IgM level and axonal damage as judged by ulnar nerve CMAP amplitude at baseline were significantly associated with clinical response, it is not possible to exclude that they might have also influenced the response as possibly indicated by their relatively high, though not significant, correlation with ISS improvement. These results suggest that rituximab may be particularly effective in patients with low antibody titer and, possibly, in the early course of the disease. In previous open-label studies, a second treatment with rituximab was found effective and safe in patients with clinical worsening after an initial improvement (Pestronk et al., 2003; Renaud et al., 2006). We only performed a second administration of rituximab in one responder despite clinical stability achieving a further clinical improvement, whereas the second administration of rituximab in a nonresponder patient after 9 months resulted in a reduction of the anti-mag titer without clinical effect. Further studies on a larger number of patients are needed to assess which immunological, clinical, or neurophysiological feature may predict a better response to rituximab and whether repeated courses of the drug may be effective in patients nonresponsive to a first course with either extremely high baseline antibody titers or their insufficient reduction after therapy. References Broglio L, Lauria G (2005). Worsening after Rituximab treatment in anti-mag neuropathy. Muscle Nerve 32: Dalakas MC, Rakocevic G, Salajegheh K, Dambrosia J, Raju Raghavan, McElroy B (2006). A double-blind, placebocontrolled study of rituximab in patients with anti-mag antibody-demyelinating polyneuropathy. Ann Neurol 60:S91, S95. Hughes R, Bensa S, Willison H, Van den Bergh P, Comi G, Illa I, Nobile-Orazio E, Van Doorn P, Dalakas M, Bojar M, Swan A; the Inflammatory Neuropathy Cause and Treatment (INCAT) Group (2001). Randomized controlled trial of intravenous immunoglobulin versus oral prednilosone in chronic 106

6 inflammatory demyelinating polyradiculoneuropathy. Ann Neurol 50: Kleyweg RP, van der Mechè FGA, Schmitz PIM (1991). Interobserver agreement in the assessment of muscle strength and functional abilities in Guillain-Barrè syndrome. Muscle Nerve 14: Lunn MPT, Nobile-Orazio E (2003). Immunotherapy for IgM antimyelin-associated glycoprotein paraprotein-associated peripheral neuropathies. Cochrane Database Syst Rev 1:CD Maloney DG, Smith B, Rose A (2002). Rituximab: mechanism of action and resistance. Semin Oncol 29:2 9. Merkies I, Schmitz P, Van Der Mechè F, Van Doorn P; for the Inflammatory Neuropathy Cause and Treatment (INCAT) Group (2000). Psychometric evaluation of a new sensory scale in immune-mediated polyneuropathies. Neurology 54: Monaco S, Bonetti B, Ferrari S, Moretto G, Nardelli E, Tedesco F, Mollnes TE, Nobile-Orazio E, Manfredini E, Bonazzi L, Rizzuto N (1990). Complement-mediated demyelination in patients with IgM monoclonal gammopathy and polyneuropathy. N Engl J Med 322: Nobile-Orazio E (1998). Neuropathies associated with anti-mag antibodies and IgM monoclonal gammopathies. In: Immunological and Infectious Diseases of the Peripheral Nerve. Latov N, Wokke JHJ, Kelly JJ (Eds). Cambridge University Press, Cambridge, pp Nobile-Orazio E, Manfredini E, Carpo M, Meucci N, Monaco S, Ferrari S, Bonetti B, Cavaletti G, Gemignani F, Durelli L, Barbieri S, Allaria S, Sgarzi M, Scarlato G (1994). Frequency and clinical correlates of anti-neural IgM antibodies in neuropathy associated with IgM monoclonal gammopathy. Ann Neurol 36: Pestronk A, Florence J, Miller T, Choski R, Miller T, Choksi R, Al-Lozi MT, Levine TD (2003). Treatment of IgM antibody associated polyneuropathies using rituximab. J Neurol Neurosurg Psychiatr 74: Renaud S, Gregor M, Fuhr P, Lorenz D, Deuschl G, Gratwohl A, Steck AJ (2003). Rituximab in the treatment of polyneuropathy associated with anti-mag antibodies. Muscle Nerve 27: Renaud S, Fuhr P, Gregor M, Schweikert K, Lorenz D, Daniels C, Deuschl G, Gratwohl A, Steck AJ (2006). High-dose rituximab and anti-mag-associated polyneuropathy. Neurology 66: