10 M.R. Farlow et al. / Treatment of Moderately Severe to Severe Alzheimer’s Disease 627 By week 13, those in the 20 g arm signals were giving a treatment difference of greater than 4.0 679 628 observed to show an increase in mean (SEM) SIB of points (difference (80% CI) = 4.09 (1.33, 6.85), 680 629 1.16 (1.15) from baseline, while the placebo mean p = 0.029). Similarly, including patients who dropped 681 630 (SEM) SIB decreased by –0.79 (1.33) points from out prior to receiving all 7 doses (i.e., the FAS), the 682 631 baseline during this same time period (difference 15-week mean (SEM) SIB increased from baseline by 683 632 [80% CI] = 1.94 [–0.31, 4.19], p = 0.134). At the 5- 1.77 (1.34) points in the 20 g arm, while the placebo 684 633 week secondary endpoint, comparison of the bryo- group showed a decline in mean (SEM) SIB = –1.82 685 634 statin 20 g cohort versus the placebo cohort favored (1.73), giving a treatment difference favoring bryo- 686 f 635 bryostatin (difference [80% CI] = 2.96 [0.58, 5.34], statin (80% CI) = 3.59 (0.79, 6.39), (p = 0.050). 687 roo 636 p = 0.056). Among secondary outcome measures, we found 688 637 Among patients exposed to the complete dosing that the ADCS-ADL-SIV mean score favored bryo- 689 638 regimen and who had a 13-week SIB assessment statin 20 g versus placebo at week 13 (p = 0.082) 690 639 (CAS patient group), evidence of benefit in the SIB in the CAS patient subset. In contrast, there were 691 640 scores favored the 20 g bryostatin arm versus the no benefits of bryostatin 20 g versus placebo in 692 rP 641 placebo arm for the primary SIB endpoint at weeks 5 the ADCS-ADL-SIV mean score at week 13 from 693 642 and 15 (Fig. 2B; Table 2). Baseline SIB scores were baseline in the FAS (p = 0.104). 694 643 similar across all three treatment arms. Much smaller effects were seen between these two 695 644 At week 13, the mean (SEM) SIB increased by arms in the ADCS-ADL-SIV scores at week 5 and 696 1.51 (1.12) points from baseline in the 20 g arm, tho 645 week 9. The NPI changes from baseline did not show 697 646 while placebo patients showed a decrease in their convincing improvement in both bryostatin dosing 698 647 mean (SEM) SIB scores from baseline of –1.12 (1.39) arms versus the placebo patients. 699 648 (difference [80% CI] = 2.63 [0.35, 4.91], p = 0.070). 649 At week 5, there was also evidence of benefit in Exploratory analyses 700 mean SIB scores from baseline among 20 g bryo- Au 650 651 statin patients versus the placebo patients (difference Because PKC isozymes have been shown in pre- 701 652 [80% CI] = 4.00 [1.63, 6.38], p = 0.016). See Table 2. vious pre-clinical studies to regulate the NMDA 702 653 No differences in mean SIB changes from baseline receptor (see Discussion below), the clinical effects 703 654 in the CAS patient sub-group were seen in the 40 g of a known blocker of the NMDA receptor, meman- 704 treatment arm versus the placebo arm at any follow- tine, used as a baseline SOC during the bryostatin d 655 705 656 up time point (see Fig. 2B). Similarly, pooling the protocols, were considered particularly important 706 cte 657 bryostatin-treated patients across both dosing arms among the pre-specified exploratory parameters. As 707 658 did not produce statistically significant differences in mentioned above (see Statistics), except for the multi- 708 659 mean SIB changes compared to the placebos at any variate Rank Sum test, all p-values for the exploratory 709 660 follow-up time points. analyses are reported as 2-sided, and an alpha level 710 661 As seen in Fig. 3, the dropout rate was very similar of 0.05 was considered statistically significant. 711 rre 662 between the 20 g arm and the placebo arm through- As can be seen in Fig. 5A, the patients who received 712 663 out the course of the trial. In contrast, the dropout no concomitant memantine baseline therapy in the 713 664 rate was very high for the 40 g arm as compared 20 g bryostatin arm showed evidence of a sustained 714 665 to the other two arms. The high dropout rate among SIB improvement from baseline over the course of 715 co 666 those exposed to the highest dose of bryostatin is most the trial. In contrast, the patients who did receive 716 667 likely the result of more side effects and AEs experi- baseline memantine (Fig. 5B) showed no evidence 717 668 enced by patients in this arm and is consistent with a of SIB improvement over time. Among the patients 718 669 lack of PKC activation efficacy at this dosing level. not receiving memantine, the mean SIB change at 719 Un 670 When we assessed post dosing outcomes on the the average of week 13 and week 15 time points 720 671 SIB, we found that patients in the 20 g arm showed from baseline was significantly greater in the 20 g 721 672 benefits from baseline at week 15 in both the CAS bryostatin arm as compared to placebo patients (dif- 722 673 and FAS patient groups (See Fig. 4). Among patients ference (95% CI) = 6.1 (1.5, 10.7) points; p = 0.012). 723 674 who were exposed to the complete dosing regi- This improvement of SIB scores persisted after con- 724 675 men (Completers), the 15-week mean (SEM) SIB trolling for baseline SIB and MMSE-2 strata at 725 676 increased from baseline by 1.96 (1.23) points in randomization in Analysis of Covariance models. For 726 677 the 20 g treatment group, while the placebo group comparison purposes, we also considered the change 727 678 showed a decline in mean (SEM) SIB = –2.13 (1.76), in SIB at week 13 from baseline, the original primary 728 M.R. Farlow et al. / Treatment of Moderately Severe to Severe Alzheimer’s Disease 11 Table 2 Results of the MMRM analysis for the completer analysis set Week 5 Week 9 Week 13 Diff. 20 g versus Placebo (80% CI) 4.0 (1.6, 6.4) 1.9 (–0.6, 4.3) 2.6 (0.4, 4.9) One-sided p-value 0.016 0.165 0.070 Diff. 40 g versus Placebo (80% CI) 2.1 (–0.4, 4.6) 0.1 (–2.3, 2.5) 1.5 (–0.7, 3.8) One-sided p-value 0.137 0.476 0.191 f Table 3 roo Results of the MMRM analysis for the full analysis set Week 5 Week 9 Week 13 Diff. 20 g versus Placebo (80% CI) 3.0 (0.6, 5.3) 1.0 (–1.4, 3.4) 1.9 (–0.3, 3.4) One-sided p-value 0.056 0.290 0.134 Diff. 20 g versus Placebo (80% CI) 0.6 (–1.7, 3.0) –0.6 (–2.9, 1.7) 0.8 (–1.4, 3.0) rP One-sided p-value 0.368 0.638 0.314 tho d Au cte Fig. 4. Improvement scores in SIB scores – through week 15. SIB improves throughout the trial, with a decline in SIB for the placebo rre patients (lower curve). 729 endpoint, among patients who received no concomi- 20 g bryostatin cohort also showed significant bryo- 744 730 tant baseline memantine therapy. Results of this statin benefit (p < 0.024), 2-tailed, p < 0.05. 745 731 analysis were shown to produce significant improve- Finally, in the trend analyses, we found that the 746 co 732 ment in the mean SIB change from baseline in the SIB values did not increase over time for the placebo 747 733 20 g versus the placebo treatment arm (difference patients under the MMRM models, resulting in slopes 748 734 (95% CI) = 5.6 (0.4, 10.9) points; p = 0.035). that were non-significantly different from zero (e.g., 749 In the analysis using the method of Wei and Lachin ‘zero-slopes’). In contrast, the SIB slopes for the Un 735 750 736 [27] that simultaneously tested for the treatment dif- 20 g bryostatin patients who did not receive baseline 751 737 ferences in SIB at week 5, week 9, and week 13 memantine were found to be statistically signifi- 752 738 from baseline, we found results that were consis- cant, giving a slope (95% CI) = 0.38 (0.18, 0.57) SIB 753 739 tent with those of the univariate analyses shown points per week in the random intercept model, and 754 740 above (see Table 4). These sustained positive results a slope (95% CI) = 0.38 (0.18, 0.59) points per week 755 741 over time provide evidence on the superiority of the in the random intercept and slope model. The inter- 756 742 treatment over placebo. A pre-specified ANCOVA action terms, which indicate a difference in treatment 757 743 analysis for the interaction of memantine with the effect by arm, were significant in both mixed effects 758 12 M.R. Farlow et al. / Treatment of Moderately Severe to Severe Alzheimer’s Disease f roo rP tho Fig. 5. SIB improvement signals (A) are clear with repeated doses of bryostatin in the absence of memantine. No such improvement was apparent with SOC memantine (B). Table 4 Pre-specified exploratory analytical results among memantine free patients Au SIB Change from Baseline Placebo 20 g Bryostatin T-test Delta Mean (SD) Delta Mean (SD) t-statistic (p val) Week 5 –1.20 (10.26) 3.44 (5.75) 1.56 (0.134) Week 9 0.79 (7.44) 3.47 (7.04) 1.00 (0.329) Week 13 –1.14 (6.89) 4.50 (7.01) 2.22 (0.035) d Week 13/15 –0.68 (6.71) 5.41 (5.43) 2.71 (0.012) Wei-Lachin test17 T-test Wilcoxon Test cte Weight Z (1-sided p-value) Z (1-sided p-value) Equal 1.63 (0.052) 1.77 (0.039) Variance−1 1.55 (0.060) 1.75 (0.040) Optimal 1.10 (0.135) 1.63 (0.052) Trend Analysis Random Intercept Model Random Intercept, Slope Model rre Slope (Placebo) (95% CI) 0.019 (–0.19, 0.22) 0.019 (–0.19, 0.22) Slope (Bryostatin) (95% CI) 0.38∗∗ (0.18, 0.56) 0.38∗∗ (0.18, 0.57) Interaction (95% CI) 0.36∗ (0.08, 0.64) 0.36∗ (0.08, 0.64) ∗ p < 0.05; ∗∗ p < 0.001; 2-tailed t-test, alpha of 0.05 for significance. co 759 models (p < 0.012, see Table 4). Trends of individual As described above, the memantine naı̈ve patients 770 760 SIB scores over time from baseline of the patients in the 20 g bryostatin arm showed evidence of sus- 771 761 who received 20 g bryostatin (20 g bryostatin, tained benefit of SIB improvement from baseline over 772 Un 762 memantine-free; and placebo with memantine-free) the course of the trial. This evidence was also appar- 773 763 are illustrated in Fig. 6. The treatment SIB trend was ent for the patients in the unadjusted FAS (or mITT) 774 764 highly statistically different from 0 (p < 0.001) for and more apparent for the unadjusted memantine 775 765 the 20 g arm, while the treatment SIB trend was not free patients (Fig. 7A, B). An adjusted mixed-effects 776 766 statistically different from 0 for the placebo arm. The model incorporating time as a factor variable and time 777 767 trends for individual patients are illustrated in Fig. 6, by treatment interactions at each time point produced 778 768 for patients off memantine, both in response to the estimates close in value to the unadjusted mean SIB 779 769 20 g protocol and in response to placebo. scores from baseline (Fig. 7A-C). 780 M.R. Farlow et al. / Treatment of Moderately Severe to Severe Alzheimer’s Disease 13 f roo rP tho d Au cte rre co Fig. 6. Individual SIB slopes (e.g., trends over time) from baseline (various color lines), and overall treatment SIB slopes (darker black lines) for memantine-free 20 g bryostatin arm, (Top); and memantine-free, placebo arm, (Bottom), respectively. Based on the statistical analysis, only the 20 g bryostatin, memantine-free group, overall treatment (dark black line, Top) shows a significant (p < 0.001) positive SIB trend Un (SIB improvement with repeated doses over time) suggesting a treatment effect of bryostatin for this group only. With memantine present, neither the 20 g bryostatin arm nor the placebo arm showed a significant positive SIB trend. 781 DISCUSSION improvement signals could only be observed in the 786 Completer populations for the primary data analysis 787 782 Bryostatin 20 g did not meet pre-specified out- and not at a commonly accepted level of statistical 788 783 come criteria, but planned analyses showed what we significance. However, these SIB improvement sig- 789 784 believe are consistent signals of benefit for the drug at nals could be observed at 15 weeks, i.e., four weeks 790 785 this dose. We would emphasize, however, that these after the termination of the dosing protocol at week 11 791 14 M.R. Farlow et al. / Treatment of Moderately Severe to Severe Alzheimer’s Disease f roo rP tho Au Fig. 7. A) Mean SIB changes (unadjusted) from baseline for the FAS subset. 80% confidence intervals are given in Table 3. B) Mean SIB changes (unadjusted) from baseline for the FAS subset for patients not on memantine. C) Mean SIB changes (adjusted) from baseline for the FAS subset for patients not on Memantine. Error Bars = SEM. d 792 for both the FAS and CAS subsets. One exploratory It is also worth re-emphasizing that the meman- 815 analysis that was pre-specified (the ANCOVA for tine, often used for symptomatic relief, here blocked cte 793 816 794 memantine) and the other post-hoc exploratory anal- all signals of bryostatin induced SIB improvement. 817 795 yses, however, did reach significance at the 2-tailed, Chronic memantine drug therapy has not been shown 818 796 p < 0.05 level with multiple analytic tests such as a to have lasting benefit. Testing for the interaction 819 797 Trend Analysis and Wei-Lachin integrated measure- of memantine baseline therapy with bryostatin effi- 820 rre 798 ments (Fig. 6A, B). These exploratory analytic results cacy was pre-specified (in the Statistical Analysis 821 799 can guide further clinical trials that will use the 20 g Plan) when the data were still blinded, prior to 822 800 dose on patients who are not on concomitant baseline unblinding and data analysis. For the effective 20 g 823 801 memantine therapy. dose, in the absence of baseline memantine, only 1 824 In the present report, the 40 g produced little or in 16 patients showed a SIB decline for the week co 802 825 803 no benefit at the frequency of administration in the 13–week 15-week endpoint. In contrast, 9 in 22 826 804 selected protocol design. Nor would benefit from this patients in the patient group receiving 40 g bryo- 827 805 40 g dose, at the frequency administered—based statin while on memantine, and 20 in 36 placebo 828 Un 806 on prior pre-clinical and Compassionate Use trial patients showed a SIB decline. The principle targets 829 807 experience—be expected to be effective. However, of bryostatin, PKC isozymes, are known to regu- 830 808 this higher dose did provide a dose-limit for future late NMDA receptor functions, which are blocked 831 809 trials as well as context for a lower dose, 20 g, at by memantine. Therefore, it is not surprising that 832 810 this frequency. Because this was a first-in-humans the blockade of the NMDA receptor could off- 833 811 multiple dose trial, there was no way to know a pri- set most if not all the bryostatin treatment effect. 834 812 ori how the pre-clinical dosing data would translate PKC regulation of the NMDA receptor functions 835 813 into human dosing until we conducted the present includes increasing NMDA conductance by relieving 836 814 exploratory trial. Mg++ blockade, controlling trafficking of the NMDA 837 M.R. Farlow et al. / Treatment of Moderately Severe to Severe Alzheimer’s Disease 15 838 receptor to the neuronal membranes, and enhancing REFERENCES 885 839 NMDA-induced synaptogenesis. This synaptogene- 840 sis, a primary mechanism of action of bryostatin [1] Weiner HL, Frenkel D (2006) Immunology and 886 immunotherapy of Alzheimer’s disease. Nat Rev Immunol 887 841 demonstrated in a variety of pre-clinical models, is 6, 404-416. 888 842 mediated by bryostatin-PKC epsilon enhancement of [2] Reisberg B, Doody R, Stöffler A, Schmitt F, Ferris S, 889 843 several synaptic growth factors that include BDNF, Möbius HJ; Memantine Study Group (2003) Memantine 890 844 NGF, and IGF. in moderate-to-severe Alzheimer’s disease. N Engl J Med 891 348, 1333-1341. 892 845 While the memantine interaction with bryostatin [3] Farlow MR, Salloway S, Tariot PN, Yardley J, Moline ML, f 893 846 adds complexity to the potential benefit of bryostatin Wang Q, Brand-Schieber E, Zou H, Hsu T, Satlin A (2010) 894 roo 847 for AD patients, we would submit that it also provides Effectiveness and tolerability of high-dose (23 mg/d) versus 895 848 some additional evidence for this potential benefit. standard-dose (10 mg/d) donepezil in moderate to severe 896 Alzheimer’s disease: A 24-week, randomized, double-blind 897 849 Namely, an effect of bryostatin that occurred only by study. Clin Ther 32, 1234-1251. 898 850 chance would not be likely to be eliminated entirely [4] Farlow M, Veloso F, Moline M, Yardley J, Brand-Schieber 899 851 in only patients who received SOC memantine (see E, Bibbiani F, Zou H, Hsu T, Satlin A (2011) Safety 900 rP 852 Fig. 5A, B). and tolerability of donepezil 23 mg in moderate to severe 901 Alzheimer’s disease. BMC Neurol 11, 57. 902 853 The apparent persistence of the bryostatin-induced [5] Farlow MR (2005) Moderate to severe Alzheimer’s disease: 903 854 SIB improvement signals is consistent with a long- Definition and clinical relevance. Neurology 65, S1-S4. 904 855 lasting consequence of PKC epsilon-growth factor [6] Terry RD, Masliah E, Salmon DP, Butters N, DeTeresa R, 905 Hill R, Hansen LA, Katzman R (1991) Physical basis of 906 tho 856 effects that could induce the growth and/or mat- cognitive alterations in Alzheimer’s disease: Synapse loss 907 857 uration of synaptic networks in the brain. This is the major correlate of cognitive impairment. Ann Neurol 908 858 might translate into long-lasting benefit in cognitive 30, 572-580. 909 859 function. [7] Scheff SW, Price DA (2006) Alzheimer’s disease-related 910 alterations in synaptic density: Neocortex and hippocampus. 911 860 Although the analyses of the primary endpoint J Alzheimers Dis 9(3 Suppl), 101-115. 912 Au 861 at 13 weeks was not significant for the full data [8] Scheff SW, Price DA, Schmitt FA, Mufson EJ (2006) Hip- 913 862 set (FAS), the data did provide evidence of bryo- pocampal synaptic loss in early Alzheimer’s disease and 914 863 statin’s improvement signals of the SIB scores at mild cognitive impairment. Neurobiol Aging 27, 1372-1384. 915 [9] Coleman P, Federoff H, Kurlan R (2004) A focus on the 916 864 13 weeks for the Completers Set and at 15 weeks, synapse for neuroprotection in Alzheimer disease and other 917 865 30 days after drug dosing completion. Pre-specified dementias. Neurology 63, 1155-1162. 918 exploratory analyses, moreover, although, in many d 866 [10] Hongpaisan J, Sun MK, Alkon D (2011) PKC Activation 919 867 cases, implemented in a post-hoc framework, did pro- prevents synaptic loss, A elevation, and cognitive deficits 920 in Alzheimer’s disease transgenic mice. J Neurosci 31, 630- 921 cte 868 vide evidence of significant benefit throughout the 643. 922 869 lower dose (20 g) protocol. The totality of these [11] Sun MK, Hongpaisan J, Nelson TJ, Alkon DL (2008) Post- 923 870 analyses, therefore, suggest that the trial showed evi- stroke neuronal rescue and synaptogenesis mediated in vivo 924 871 dence of bryostatin’s SIB improvement signals, in by PKC in adult brains. Proc Natl Acad Sci U S A 105, 925 13620-13625. 926 872 the absence of baseline memantine, that warrant fur- [12] Nelson TJ, Cui C, Luo Y, Alkon DL (2009) Reduction of 927 rre 873 ther trials to evaluate bryostatin’s potential utility to beta-amyloid levels by novel PKC (epsilon) activators. J 928 874 improve cognitive function(s) as well as to provide Biol Chem 284, 34514-34521. 929 875 symptomatic relief and/or to delay cognitive decline [13] Sen A, Nelson TJ, Alkon DL, Hongpaisan J (2018) Loss in 930 PKC epsilon causes downregulation of MnSOD and BDNF 931 876 of patients with moderately severe to severe AD. expression in neurons of Alzheimer’s disease hippocampus. 932 co J Alzheimers Dis 63, 1173-1189. 933 [14] Sun MK, Hongpaisan J, Alkon DL (2009) Post-ischemic 934 877 ACKNOWLEDGMENTS PKC activation rescues retrograde and anterograde long- 935 term memory. Proc Natl Acad Sci U S A 106, 14676-14680. 936 This study was sponsored by Neurotrope, Inc. [15] Sen A, Alkon DL, Nelson TJ (2012) Apolipoprotein E3 937 Un 878 (ApoE3) but not ApoE4 protects against synaptic loss 938 879 Authors’ disclosures available online (https:// through increased expression of protein kinase C epsilon. 939 880 www.j-alz.com/manuscript-disclosures/18-0759r3). J Biol Chem 287, 15947-15948. 940 [16] Alkon DL, Sun MK, Nelson TJ (2007) PKC signal- 941 ing deficits: A mechanistic hypothesis for the origins of 942 881 SUPPLEMENTARY MATERIAL Alzheimer’s disease. Trends Pharmacol Sci 28, 51-60. 943 [17] Sun MK, Nelson TJ, Alkon DL (2015) Towards universal 944 882 The supplementary material is available in therapeutics for memory disorders. Trends Pharmacol Sci 945 36, 384-394. 946 883 the electronic version of this article: http:// [18] Nelson TJ, Sun MK, Lim C, Sen A, Khan T, Chirila FV, 947 884 dx.doi.org/10.3233/JAD-180759. Alkon DL (2017) Bryostatin effects on cognitive function 948 16 M.R. Farlow et al. / Treatment of Moderately Severe to Severe Alzheimer’s Disease 949 and PKC in Alzheimer’s phase IIa and expanded access [23] Guy W (1976) Clinical Global Impressions. ECDEU 968 950 trials. J Alzheimers Dis 58, 521-535. Assessment Manual for Psychopharmacology—Revised. 969 951 [19] Folstein MF, Folstein SE, McHugh PR (1975) Mini-Mental U.S. Department of Health, Education, and Welfare; Public 970 952 State: A practical method for grading the cognitive state of Health Service, Alcohol; Drug Abuse, and Mental Health 971 953 patients for clinician. J Psychiatr Res 12, 189-198. Administration; National Institute of Mental Health; Psy- 972 954 [20] Varterasian ML, Pemberton PA, Hulburd K, Rodriguez DH, chopharmacology Research Branch; Division of Extramural 973 955 Murgo A, Al-Katib AM (2001) Phase II study of bryostatin Research Programs, Rockville, MD, pp. 218-222. 974 956 1 in patients with relapsed multiple myeloma. Invest New [24] Cummings JL (1997) The Neuropsychiatric Inventory. Neu- 975 957 Drugs 19, 245-247. rology 48(5 Suppl 6), 10S-16S. 976 958 [21] Schmitt FA, Ashford W, Emesto C, Saxton J, Schneider LS, [25] Wilcoxon F (1945) Individual comparisons by ranking 977 f 959 Clark CM, Ferris SH, Mackell JA, Schafer K, Thal LJ (1997) methods. Biometrics 1, 80-83. 978 roo 960 The severe impairment battery: Concurrent validity and the [26] Rubinstein L, Crowley J, Ivy P, LeBlanc M, Sargent D 979 961 assessment of longitudinal change in Alzheimer’s disease. (2009) Randomized phase II designs. Clin Cancer Res 15, 980 962 The Alzheimer’s disease cooperative study. Alzheimer Dis 1883-1890. 981 963 Assoc Disord 11, 51-56. [27] Wei LJ, Lachin JM (1984) Two sample asymptotically dis- 982 964 [22] Galasko D, Bennett D, Sano M, Ernesto C, Thomas R, tribution free tests for incomplete multivariate observations. 983 965 Grundman M, Ferris S (1997) An inventory to assess activi- J Am Stat Assoc 79, 653-661. 984 rP 966 ties of daily living for clinical trials in Alzheimer’s disease. 967 Alzheimer Dis Assoc Disord 11(Suppl 2), S33-39. tho d Au cte rre co Un
Enter the password to open this PDF file:
-
-
-
-
-
-
-
-
-
-
-
-