1. Genomic Guided Therapy in Myeloma
Dr. A. Keith Stewart
Vasek and Anna Maria Professor of Cancer Research
Carlson-Nelson Endowed Director, Center for Individualized Medicine
Mayo Clinic

2. Improving response rates and CR with IMiD
and Proteasome inhibitor combination therapies 2

3. Improving Survival in MM
25% of patients live
less than 3 years

4. mSMART 2.0 Variable Outcomes in MM
High 20%
Intermediate 20%
Standard 60%
FISH
Del 17p
t(14;16)
t(14;20)
GEP
High risk signature
Others
Hyperdiploid
t(11;14)
t(6;14)
FISH
t(4;14)
Cytogenetic deletion 13 or hypodiploidy
PCLI ≥3%
3 years
4-5 years
10 years

5. Proteasome Inhibitor Resistant MM Progenitor Cells
Todays Goal: To Describe Three Clinically Relevant, Multiple Myeloma Drug Resistance Mechanisms and Their Impact on Treatment Decision Making
Clonal Tides
Cereblon and IMiDs
Proteasome Inhibitor Resistant MM Progenitor Cells
Genomic Monitoring

6. Clinical Course of t(4;14) High Risk Patient

7. Clonal Tides 5 unique clones at diagnosis
Variable chemotherapy response
Minor drug resistant clone lethal

8. Implications Multiple clones with variable drug sensitivity
Combination chemotherapy a necessity
Resuscitation of drug sensitive clones
Once resistant not always resistant
Continuous suppressive therapy logical
Minor drug resistant clones lethal
Need to understand mechanism of resistance as a means to eradicate

9. Relapsed Multiple Myeloma:
Carfilzomib, Lenalidomide, and Dexamethasone vs Lenalidomide and Dexamethasone in Patients with
Relapsed Multiple Myeloma:
Interim Results from ASPIRE, a Randomized, Open-Label, Multicenter Phase 3 Study
A. Keith Stewart, S. Vincent Rajkumar, Meletios A. Dimopoulos, Tamás Masszi, Ivan Spicka, Albert Oriol, Roman Hájek, Laura Rosiñol, David S. Siegel, Georgi G. Mihaylov, Vesselina Goranova-Marinova, Péter Rajnics, Aleksandr Suvorov, Ruben Niesvizky, Andrzej Jakubowiak, Jesus F. San Miguel, Heinz Ludwig, Naseem Zojwalla, Margaret E. Tonda, Biao Xing, Philippe Moreau and Antonio Palumbo
NEJM, Jan 8th, 2015

10. ASPIRE Study Design KRd Rd Randomization 28-day cycles N=792
Category (XX)
4/20/2017
ASPIRE Study Design
28-day cycles
Randomization
N=792
Stratification:
β2-microglobulin
Prior bortezomib
Prior lenalidomide
KRd
Carfilzomib 27 mg/m2 IV (10 min)
Days 1, 2, 8, 9, 15, 16 (20 mg/m2 days 1, 2, cycle 1 only)
Lenalidomide 25 mg Days 1–21
Dexamethasone 40 mg Days 1, 8, 15, 22
After cycle 12, carfilzomib given on days 1, 2, 15, 16
After cycle 18, carfilzomib discontinued Rd
Lenalidomide 25 mg Days 1–21
Dexamethasone 40 mg Days 1, 8, 15, 22

11. Secondary Endpoints: Response
Percentage of Patients
sCR
14.1% vs 4.3%
P<.0001
Median duration of response was 28.6 months in the KRd group and 21.2 months in the Rd group

12. Months Since Randomization
Primary Endpoint: Progression-Free Survival ITT Population (N=792)
KRd Rd
(n=396) (n=396)
Median PFS, mo
HR (KRd/Rd) (95% CI) (0.57–0.83)
P value (one-sided) <0.0001
1.0
0.8
0.6
Proportion Surviving
Without Progression
0.4
0.2
KRd Rd
The cutoff date for the interim analysis was June 16, 2014
In the carfilzomib and control groups, 118 (29.8%) and 86 (21.7%) patients, respectively, were still receiving study treatment
At the time of the prespecified interim analysis, 431 progression-free survival events were observed
The study met its primary objective of demonstrating that carfilzomib improves progression-free survival when administered with lenalidomide and dexamethasone
With an estimated hazard ratio of (95% CI, to 0.834), the P value (P<0.0001) crossed the prespecified stopping boundary
The primary end point was evaluated using a group sequential design with one interim analysis
In total, 526 progression-free survival events were needed to provide 90% power to detect a 25% reduction in risk of disease progression or death (hazard ratio of 0.75) at a one-sided significance level of 0.025
The interim analysis was to be performed when approximately 420 progression-free survival events (80% of the planned total) were observed
An O’Brien–Fleming type of efficacy stopping boundary was calculated using the Lan–DeMets alpha spending function approach based on the number of events observed at the data cutoff date
0.0 6 12 18 24 30 36 42 48
Months Since Randomization
No. at Risk:
KRd Rd

13. Months Since Randomization
Primary Endpoint: Progression-Free Survival ITT Population (N=792)
KRd Rd
(n=396) (n=396)
Median PFS, mo
HR (KRd/Rd) (95% CI) (0.57–0.83)
P value (one-sided) <0.0001
1.0
0.8
0.6
Proportion Surviving
Without Progression
0.4
0.2
KRd Rd
The cutoff date for the interim analysis was June 16, 2014
In the carfilzomib and control groups, 118 (29.8%) and 86 (21.7%) patients, respectively, were still receiving study treatment
At the time of the prespecified interim analysis, 431 progression-free survival events were observed
The study met its primary objective of demonstrating that carfilzomib improves progression-free survival when administered with lenalidomide and dexamethasone
With an estimated hazard ratio of (95% CI, to 0.834), the P value (P<0.0001) crossed the prespecified stopping boundary
The primary end point was evaluated using a group sequential design with one interim analysis
In total, 526 progression-free survival events were needed to provide 90% power to detect a 25% reduction in risk of disease progression or death (hazard ratio of 0.75) at a one-sided significance level of 0.025
The interim analysis was to be performed when approximately 420 progression-free survival events (80% of the planned total) were observed
An O’Brien–Fleming type of efficacy stopping boundary was calculated using the Lan–DeMets alpha spending function approach based on the number of events observed at the data cutoff date
0.0 6 12 18 24 30 36 42 48
Months Since Randomization
No. at Risk:
KRd Rd

14. PFS by Risk Group KRd (n=396) Rd Risk Group by FISH N Median, monthsHR
P-value
(one-sided)
High 48
23.1 52
13.9
0.70
0.083
Subgroup analysis of PFS as determined by IRC; ITT population

15. PFS by Risk Group KRd (n=396) Rd Risk Group by FISH N Median, monthsHR
P-value (one-sided)
High 48
23.1 52
13.9
0.70
0.083
Standard
147
29.6
170
19.5
0.66
0.004
Subgroup analysis of PFS as determined by IRC; ITT population

16. Months Since Randomization
Secondary Endpoints: Interim Overall Survival Analysis Median Follow-Up 32 Months
1.0
KRd Rd
(n=396) (n=396)
Median OS, mo NE NE
HR (KRd/Rd) (95% CI) (0.63–0.99)
P value (one-sided)
0.8
0.6
Proportion Surviving
0.4
0.2
KRd Rd
0.0
Since the primary objective was met, an interim analysis of overall survival was carried out.
Using the same cutoff date, 305 events had occurred (60% of the prespecified 510 events required for final analysis)
Median follow-up was 32.3 and 31.5 months in the carfilzomib and control groups
Median overall survival was not reached in either group, with a hazard ratio of (95% CI, to 0.985; P=0.0182) trending in favor of the carfilzomib group
However, these results did not cross the prespecified stopping boundary (P=0.005) for overall survival at the interim analysis
The Kaplan–Meier 24-month overall survival rates were 73.3% (95% CI, 68.6 to 77.5) and 65.0% (95% CI, 59.9 to 69.5) in the carfilzomib and control groups, respectively
The unadjusted P value from the stratified log-rank test comparing the overall survival curves up to 2 years was 6 12 18 24 30 36 42 48
Months Since Randomization
No. at Risk:
KRd Rd
Median OS was not reached; results did not cross the prespecified stopping boundary (P=0.005) at the interim analysis

17. Health-Related Quality-of-Life70
Carfilzomib group
Control group
EORTC Global Health Status improved in the KRd group vs the Rd group over 18 cycles of treatment (P=0.0001) 65
EORTC QLQ-C30 Global Health Status/Quality-of-Life Score 60
The minimal important difference for between-group differences on the QLQ-C30 Global Health Status/Quality of Life scale is 5 points, which was met at cycle 12 (5.56 points) and approached at cycle 18 (4.81 points) 55 50
Cycle 1
(Baseline)
Cycle 3
Cycle 6
Cycle 12
Cycle 18
Assessment Time Point (Day 1)

18. Conclusions
PFS was significantly improved by 8.7 months with KRd (HR, 0.69; P<0.0001)
An unprecedented median PFS of 26.3 months with KRd
Interim OS analysis: trend in OS favoring the KRd group; Kaplan-Meier 24-month OS rates 73.3% (KRd) versus % (Rd)
ORR was higher with KRd (87.1% vs 66.7%); significantly more patients achieved ≥CR (31.8% vs 9.3%)
QoL Global Health Status improved

19. Inferred Conclusions
Combination of drugs is better
Longer duration of combination therapy may have helped
High Risk disease is still worse
Resistance eventually emerges

20. IMiD Structures
Side effects
Potency
Potency

21. Cereblon
Cereblon on chromosome 3 was first described as associated with human intelligence (Cerebral protein with Lon protease)
Functions in the brain as an ionic channel regulator
Highly conserved from plants to humans, broadly expressed
Forms an E3ligase complex with DDB1, Cul4A, Roc1

22. Cereblon Levels are Highest in MM, Leukemias, and Neuroblastoma

23. Lenalidomide Resistant MM Cells Lack Cereblon
MM1.S
MM1.S res
CRBN
b-actin
Zhu YX, et al. Blood. 2011;118:

24. CRBN expression as a percentage of the mean levels in all MM
Gene Expression Levels of Cereblon Predict Response Rate to Pomalidomide
33%
19%
CRBN expression as a percentage of the mean levels in all MM 0%
N =
Schuster SR, et al. Leuk Res. Jan 2014; 38(1):23-28

25. Gene Expression Levels of Cereblon Predict Overall Survival of Pomalidomide Treated Patients
9.1 vs 27 months
Schuster SR, et al. Leuk Res. Jan 2014; 38(1):23-28

26. CRBN Binding Proteins Altered by Lenalidomide
Zhu et al. Blood Jun 9. [Epub ahead of print]

27. CRBN Binding Proteins Altered by Lenalidomide
Zhu et al. Blood Jun 9. [Epub ahead of print]

28. CRBN Binding Proteins Altered by Lenalidomide
Zhu et al. Blood Jun 9. [Epub ahead of print]

29. IKZF1/IRF4/Myc Degradation After Lenalidomide
Time h h h h
Len
IKZF1
IKZF3
IRF4
MYC
b-actin

31. CRBN/IZKF1/IRF4 Expression and Drug Resistance
XG MM1.S KMS H JJN EJM OCI-MY FR SKMM
IZKF1
IZKF3
b-actin
Len Sensitive
Len Resistant
CRBN
IRF4
IZKF1 L208R
damaging mutation
t(6;14) IgH-IRF4 translocation
Zhu et al. Blood Jun 9. [Epub ahead of print]

32. Degradation of Ikaros by Cereblon Binding Small Molecules
Stable myeloma cell line 8226 expressing IKZF1 and luciferase fusion gene treated with Imids for 24 hours

33. Sensitivity of MM to Lenalidomide is Correlated with IKAROS Degradation Efficiency
Most resistant cell lines
Most Sensitive cell lines
MM cell lines with adenoviral vector expressing IKAROS 1 - luciferase fusion gene. Luciferase activity was measured and normalized to cells treated with DMSO

34. Proteasome Inhibitors block IKAROS degradation by Lenalidomide

35. Summary Cereblon is the IMiD target and accumulates with IMiD binding
Ikaros is rapidly degraded by CRBN in presence of IMiD – a process blocked by bortezomib or carfilzomib
Low Cereblon and Ikaros levels may predict response and survival
Resistance can be explained in many cases by disruption of this pathway

36. What About Proteasome Inhibitor Resistance ?

37. Druggable genome siRNA screening
Human Druggable Genome siRNA Set
Bortezomib
Day 0
siRNA - preprinted
+ Transfection reagent
+ Cells / fresh media
Day 4
Assess viability
During screening we used custom libraries of unique siRNA to systematically silence, one by one, over 7,000 genes in myeloma cells.
On day 0 we transfected myeloma cells with siRNA, and on day 4 we assessed cellular viability.
In addition, on day 1, in some experiments, we added titrated doses of bortezomib to siRNA replicate wells, to identify genes that might modulate the activity of this drug
CellTitre Glo
Incubate
In bortezomib

38. Synthetic Lethal siRNA Screens Identify IRE1 as Essential for Bortezomib Cytotoxicity in MM Cells
Kinome
(650 genes)
← Sensitizing kinases (on siRNA knockdown)
Resistance kinases (on siRNA knockdown) →
Bortezomib
IRE1
Druggable Genome
(7000 genes)
← Sensitizing genes (on knockdown)
Resistance genes (on SiRNA knockdown) →
Bortezomib
IRE1
Unfolded Protein Response
IRE1
XBP1u XBP1s Plasma Cells
Tiedeman et al: Cancer Cell (2013) Sep 9; 24,

39. Proteasome Inhibitor Resistance
Loss of IRE1/XBP1s confers PI resistance
Low IRE1/XBP1s cells are pre plasmablasts
These XBPs low MM progenitors survive PI and contribute to relapse
High immunoglobulin production confers sensitivity
Tiedeman et al: Cancer Cell (2013) Sep 9; 24,

40. MM Progenitors and Role in Therapeutic Resistance
Cancer Cell (2013) Sep 9; 24,

41. A Custom Multiple Myeloma Mutation Panel for Clinical Targeted Sequencing
KM Kortüm, AK Stewart, LA Bruins, G Ahmann, G Vasmatzis, SV Rajkumar, S Kumar, A Dispenzieri, MQ Lacy, MA Gertz, R Fonseca, M Champion, PL Bergsagel and E Braggio;
Scottsdale, Arizona
Rochester, Minnesota
Jacksonville, Florida
Mayo Clinic College of Medicine
Mayo Clinic Comprehensive Cancer Center

42. Publication of 300+ MM genomes/exomes essentially defined the genetic landscape of the disease

43. Only a limited set of genes is recurrently mutated in MM
Lohr et al., Cancer Cell 2014

44. Myeloma Mutation Panel (M3P)
Recurrently mutated : FAM46C, DIS3, TP53, RB1, etc.
Actionable: BRAF, RAS, IDH1/2, FGFR3, AKT etc.
Pathways: NF-kB, MAPK, IL6-JAK-STAT, MYC, etc.
Copy Number: Hyperdiploid, del17, del13, 1q+
Drug Resistance: IMiDs, Proteasome inhibitors, Glucocorticoid
77 genes, 1271 amplicons, 271 Kb

45. Customized MM specific gene panel using semiconductor sequencing technology
Mutation and copy-number information
Low input DNA quantity (~10 ng)
Results available in clinically meaningful timeframes: 3 days from library preparation to data analysis
Low cost (~$250/sample)
Increased coverage

46. Results To Date:
142 untreated and 14 double-refractory MM patients incl. corresponding germline sample
Average sequencing depth : 600X
283 nonsynonymous mutations
Mutations found in 73% of the patients and in 66% of the panel genes

47. Top mutated genes

48. Clonal diversity of the mutational spectrum
50% of the mutations
<25%VR
Variant read frequency
23%
<10%VR
Mutations

49. Myeloma Specific Gene Mutation Panel Tracks Clonal Changes Over Time
Patient One Patient Two Patient Three

50. MAPK signaling pathway
Number of patients (n=156)
Incidence
KRAS 1
43%
NRAS
BRAF
Variant reads

51. BRAF mutations in 9% of patients
1 patient with 2 BRAF mutations
13 potentially activating mutations:
9 mutations in the activating loop, of which 6 at known actionable position V600E (4%)
4 more in the phosphate binding loop
status
protein
variant reads
untreated
p.Gly466Glu
11%
p.Gly469Arg
12%
p.Gly466Ala 5%
p.Gly464Val 8%
treated
p.Leu485Phe
31%
p.Lys499Asn
p.Asp594Asn
18%
p.Asp594Gly 7%
p.Gly596Arg
29%
p.Val600Glu
14%
44%
25%
23%
activating loop (AA )
phosphate binding loop (AA )

52. Mutations in targetable genes include IDH1 R132 mutation
Activating R132 IDH1 mutations are frequently found in glioma, but also in other malignancies including acute myeloid leukemia and have been described in myeloma.
IDH1 inhibitors are in clinical investigation and demonstrated clinical efficacy
In our 14 advanced patients we identified 2 patients (15%) with an IDH1 mutation (R132G and R132C)

53. Mutations in the Cereblon pathway in 5% of the patients
IRF4
MYC
MM cytotoxicity
CRBN
IMiDs
IKZF1/3 *
DDB1
CUL4
ROC1
CRBN 1
CUL4B
IKZF3
IRF4
In an untreated patient refractory to Len/Dex induction therapy we identified mutations in CRBN and IRF4
In a refractory patient progressed on Thalidomide, Lenalidomide and Pomalidomide IKZF3 was mutated.
4 out of 5 IRF4 mutated patients shared a K123R variant
* = gene found mutated in cohort

54. Mutation in genes associated with proteasome inhibitor resistance
XBP1 needed for plasma cell maturation and suppression of XBP1s in MM has shown to induce bortezomib resistance
In a patient that progressed on Carfilzomib we found a truncating XBP1 mutation
XBP1 (p.Leu232*)
Tumor
normal
Leung-Hagesteijn et al, Cancer Cell 2013

55. Summary Clonal heterogeneity is prevalent Minor clones can be lethal
Mutation profiling has potential to identify drug resistant clones and remission status in real time
Clinical Markers of drug response are emerging
IMiD: Cereblon, Ikaros, IRF4
Proteasome: IRE1, XBP1

56. Thankyou: Comments, Questions or Insults ?