nordicnanovector · 2016. 11. 2. · created date: 9/25/2015 1:33:58 pm

1
Comparison of murine and chimeric version of the an-CD37 anbody HH1 used for anbody-radionuclide-conjugate (ARC) therapy of non-Hodgkin Lymphoma Dahle, J. 1 , Repeo-Llamazares, A. 1 , Melhus, K.B. 1 , O’Shea, A. 1 , Generalov, R. 1 , Andersen, J.T. 2 and Heyerdahl, H. 1 1 Nordic Nanovector ASA, Kjelsåsveien 168 B, Oslo, Norway. 2 Department of Immunology, Centre for Immune Regulaon (CIR), Oslo University Hospital Rikshospitalet, Oslo, Norway. Poster presented at the Annual Congress of the European Associaon of Nuclear Medicine (EANM), Hamburg, Germany, 10–14 October 2015. INTRODUCTION The novel anbody-radionuclide-conjugate (ARC) 177 Lu-DOTA-HH1 (Betalun®) is currently in clinical phase 2 trial for treatment of non-Hodgkin Lymphoma (NHL). We are now developing a chimeric version of the HH1 anbody that can be used for mulple treatments. The HH1 anbody is a murine an-CD37 anbody and has a lower binding affinity to human Fc-receptors than a chimeric or humanised anbody and consequently a shorter biological half-life. A biological half-life similar to the half-life of the Lu-177 radionuclide, which is 6.7 days, will opmise irradiaon of the tumour and result in lower irradiaon of normal ssues. Internalisaon of a chimeric ARC in normal ssues expressing the neonatal Fc-receptor may also result in unwanted irradiaon of normal ssues. There are, however, benefits of using a chimeric anbody: 1. Reduced level and severity of human an-drug anbody responses. 2. Addional and possibly synergisc therapeuc effect from anbody dependent cellular cytotoxicity (ADCC) and complement dependent cytotoxicity (CDC). This study was performed in order to invesgate the similaries and differences of the murine and the chimeric version of the an-CD37 ARC. BINDING TO HUMAN FC g –RECEPTORS IN SILICO IMMUNOGENICITY BIODISTRIBUTION IN MICE WITH RAMOS XENOGRAFTS ACKNOWLEDGMENT ANTIBODY DEPENDENT CELLULAR CYTOTOXICITY BINDING TO CELLS IN VITRO INTERNALISATION OF ANTIBODIES IN VIVO EFFICACY OF NAKED ANTIBODIES TISSUE CROSS REACTIVITY IN VITRO CYTOTOXICITY OF ARCS CONCLUSIONS We are grateful to: Tina Bøndsdorff, Oncoinvent AS , for performing the ADCC assay, to Bergthora Eiriksdor, ARcLAS , for performing the in vivo therapy experiment, to sciensts at Covance , for performing TCR studies, to sciensts at EIR Sciences and ImmunExperts for performing in silico immunogenicity, to Sebasan Patzke, Norwegian Radium Hospital for help with microscopy and to Sylvia Kolenic, Nordic Nanovector ASA, for designing the poster. Aligned posion The immunogenicity of chHH1 and HH1 was predicted by invesgaon of the in silico interacon between 15-mer pepdes from the protein sequence of the two anbodies and the human major histocompability complex. The murine HH1 anbody was predicted to be more immunogenic than chHH1 and this was especially due to a high risk region in posion 315–340 of the heavy chain. The HH1 anbody was predicted to be of similar immunogenicity as ibritumomab while the chHH1 anbody was similar to rituximab. The red lines indicate the CDR regions of the anbodies. HH1 Tissue cross reacvity (TCR) was measured by immunohistological staining of ssue secons from human donors. TCR studies showed that both anbodies bound selecvely to lymphoid ssues. The images show binding to lymph node secons. Days since start of treatment 0 20 40 60 80 100 120 140 160 180 0,0 0,2 0,4 0,6 0,8 1,0 NaCl HH1 chHH1 SCID mice with mantle cell lymphoma xenograſts (Rec-1) were treated twice weekly with 100 m g chHH1, 100 m g HH1 or 100 m l of NaCl for four weeks (black arrows). The treatment was started the day aſter intravenous injecon of 10 million Rec-1 cells. 180 days aſter start of treatment 100 % of the mice treated with chHH1 were sll alive, while 70 % of the mice treated with HH1 were alive. All control mice were dead aſter 80 days. The difference between the chHH1 and the HH1 treatment was not stascally significant. The observed difference in efficacy is likely related to differences in inducon of ADCC and other types of anbody induced immunological toxicity. In a separate study, HH1 was found to bind more strongly than chHH1 to mouse receptor Fc g RIIb, whereas for the receptors Fc g RI and IV chHH1 bound strongly while HH1 did not bind. This difference in binding to mouse Fc-receptors correlates with the observed differences in efficacy. The chHH1 treatment was equally effecve as rituximab treatment (data not shown). chHH1 ADCC was measured using NK-cells from donor blood. HH1 did not induce anbody dependent cellular cytotoxicity in any of the tested cell lines, as compared with control, while chHH1 induced ADCC in all the tested cell lines, except for DOHH2 cells. The effector cell to target cell rao was 10:1 and the anbody concentraon was 20 m g/ml. CDC was not an acve mechanism for any of the anbodies. Urine Blood Lungs Heart Liver Spleen Kidney Stomach Small Intestine Large Intestine Ovaries Femur Muscle Lymph Node Injection site Tumor 0 10 20 30 40 50 chHH1-DOTA, CAR: 0.5 HH1-DOTA, CAR: 0.5 chHH1-DOTA, CAR: 1.7 HH1-DOTA, CAR: 0.9 chHH1-DOTA, CAR: 2.5 HH1-DOTA, CAR: 1.6 Both the murine HH1 and the chHH1 anbodies were conjugated with p-SCN-Bn-DOTA using different DOTA:Ab raos, which resulted in different CARs. Biodistribuons of the ARCs were measured 3 days aſter injecon in mice with Ramos xenograſts. All ARCs had a relevant biodistribuon in this mouse model, and the retenon in blood and uptake in normal organs and tumour were similar. There was a non-significantly higher uptake in tumours for the ARCs with lowest CAR. Ramos cells were incubated with both ARCs for 4 hours, washed and incubated further for 2 days before the viable cell concentraon was measured by flow cytometry. Specific acvity was 200 MBq/mg for both ARCs. In the figure the results for ARCs with CARs of 1.7 for the chHH1 and 0.9 for HH1 is shown. The results were similar for ARCs with lower (0.5) or higher (1.6-2.5) CAR. The two anbodies had similar affinites (Kd 2.5 nM), internalisaon and selecvity to human lymphoid ssues. chHH1 bound to all classical human Fc receptors, while HH1 bound only to human Fcg RIIa and weakly to IIb. chHH1 induced ADCC, while HH1 did not. CDC was not an acve mechanism for either of the anbodies. chHH1 was a more effecve treatment than HH1 in mice with intravenous human mantle cell lymphoma xenograſts. The uptake in normal organs and tumour xenograſts were similar for both anbodies. The cytotoxicity in vitro was similar for both anbodies. chHH1 was predicted to be less immunogenic than HH1. Control 1 2,5 5 0,0 0,2 0,4 0,6 0,8 1,0 1,2 1,4 1,6 Control chHH1, CAR: 1.7 HH1, CAR: 0.9 Control HH1 chHH1 0 10 20 30 40 50 60 70 Rec-1 DOHH2 Ramos Daudi an-CD37 anbody Luteum-177 radiaon CD37 angen CD20 angen Tumour Cell non-Hodgkin Lymphoma Betalun® – Mechanism of acon The binding of chHH1 and HH1 to Ramos cells was similar. The chelator to anbody rao (CAR) was 0.5 in this experiment. The results did not change significantly with higher CARs. The cells were incubated with the ARCs for 1 hour on ice before washing and measurement of specific binding. The binding affinity of both anbodies was 2.5 nM. HH1 0 1 2 3 4 5 0,0 0,5 1,0 1,5 2,0 2,5 3,0 Fc RI Fc RIIa-H Fc RIIa-R Fc RIIb Fc RIIIa-F Fc RIIIa-V Fc RIIIb chHH1 0 1 2 3 4 5 OD 620nm Anbody concentraon (mg/ml) Anbody concentraon (mg/ml) Binding to Fc-receptors was measured using ELISA and recombinant Fc-receptors. The chimeric anbody bound to all the classical human Fc-receptors, while the murine anbody only bound to Fc g RIIa and IIb, which are expressed on B-cells. HH1 chHH1 HH1 chHH1 rituximab The images show localisaon of HH1 and chHH1 predominantly inside Ramos cells, while rituximab is localised on the plasma membrane of the Ramos cells. The cells were incubated over night at 37 °C. Disclosure Dahle, J.: Employment, Equity ownership, Patent. Repeo-Llamazares, A.: Employment, Equity ownership. Melhus, K.B.: Employment, Equity ownership. O’Shea, A.: Employment. Generalov, R.: Employment. Andersen, J.T.: Nothing to disclose. Heyerdahl, H.: Employment, Equity ownership. 0 20 000 40 000 60 000 80 000 100 000 120 000 140 000 160 000 180 000 0 5 000 10 000 15 000 20 000 Antibody concentration (ng/ml) chHH1, CAR=0.5 HH1, CAR=0.5 Days since start of treatment Surviving fracon Anbody concentraon (ng/ml) Specific binding (number of anbodies/cell) Viable cell concentraon relave to control cells Tissue acvity (% IA/g) FcgRI FcgRlla-H FcgRlla-R FcgRllb FcgRllla-F FcgRllla-V FcgFlllb % Specific lysis Control HH1 chHH1 Immunogenicity score Anbody concentraon (mg/ml)

Upload: others

Post on 20-Aug-2020

2 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Nordicnanovector · 2016. 11. 2. · Created Date: 9/25/2015 1:33:58 PM

Comparison of murine and chimeric version of the anti-CD37 antibody HH1 used for antibody-radionuclide-conjugate (ARC) therapy of non-Hodgkin Lymphoma

Dahle, J.1, Repetto-Llamazares, A.1, Melhus, K.B.1, O’Shea, A.1, Generalov, R.1, Andersen, J.T.2 and Heyerdahl, H.1

1 Nordic Nanovector ASA, Kjelsåsveien 168 B, Oslo, Norway.2 Department of Immunology, Centre for Immune Regulation (CIR), Oslo University Hospital Rikshospitalet, Oslo, Norway.

Poster presented at the Annual Congress of the European Association of Nuclear Medicine (EANM), Hamburg, Germany, 10–14 October 2015.

INTRODUCTION

The novel antibody-radionuclide-conjugate (ARC) 177Lu-DOTA-HH1 (Betalutin®) is currently in clinical phase 2 trial for treatment of non-Hodgkin Lymphoma (NHL). We are now developing a chimeric version of the HH1 antibody that can be used for multiple treatments.

The HH1 antibody is a murine anti-CD37 antibody and has a lower binding affinity to human Fc-receptors than a chimeric or humanised antibody and consequently a shorter biological half-life. A biological half-life similar to the half-life of the Lu-177 radionuclide, which is 6.7 days, will optimise irradiation of the tumour and result in lower irradiation of normal tissues. Internalisation of a chimeric ARC in normal tissues expressing the neonatal Fc-receptor may also result in unwanted irradiation of normal tissues.

There are, however, benefits of using a chimeric antibody: 1. Reduced level and severity of human anti-drug antibody responses.2. Additional and possibly synergistic therapeutic effect from antibody dependent

cellular cytotoxicity (ADCC) and complement dependent cytotoxicity (CDC).

This study was performed in order to investigate the similarities and differences of the murine and the chimeric version of the anti-CD37 ARC.

BINDING TO HUMAN FCg–RECEPTORS

IN SILICO IMMUNOGENICITY

BIODISTRIBUTION IN MICE WITH RAMOS XENOGRAFTS

ACKNOWLEDGMENT

ANTIBODY DEPENDENT CELLULAR CYTOTOXICITY

BINDING TO CELLS IN VITRO

INTERNALISATION OF ANTIBODIES

IN VIVO EFFICACY OF NAKED ANTIBODIES

TISSUE CROSS REACTIVITY

IN VITRO CYTOTOXICITY OF ARCS

CONCLUSIONS

We are grateful to: Tina Bøndsdorff, Oncoinvent AS, for performing the ADCC assay, to Bergthora Eiriksdottir, ARticLAS, for performing the in vivo therapy experiment, to scientists at Covance, for performing TCR studies, to scientists at EIR Sciences and ImmunExperts for performing in silico immunogenicity, to Sebastian Patzke, Norwegian Radium Hospital for help with microscopy and to Sylvia Kolenic, Nordic Nanovector ASA, for designing the poster.

Aligned position

The immunogenicity of chHH1 and HH1 was predicted by investigation of the in silico interaction between 15-mer peptides from the protein sequence of the two antibodies and the human major histocompatibility complex. The murine HH1 antibody was predicted to be more immunogenic than chHH1 and this was especially due to a high risk region in position 315–340 of the heavy chain. The HH1 antibody was predicted to be of similar immunogenicity as ibritumomab while the chHH1 antibody was similar to rituximab. The red lines indicate the CDR regions of the antibodies.

HH1

Tissue cross reactivity (TCR) was measured by immunohistological staining of tissue sections from human donors. TCR studies showed that both antibodies bound selectively to lymphoid tissues. The images show binding to lymph node sections.

Days since start of treatment

0 20 40 60 80 100 120 140 160 180

Survi

ving f

ractio

n

0,0

0,2

0,4

0,6

0,8

1,0

NaClHH1chHH1

SCID mice with mantle cell lymphoma xenografts (Rec-1) were treated twice weekly with 100 mg chHH1, 100 mg HH1 or 100 ml of NaCl for four weeks (black arrows). The treatment was started the day after intravenous injection of 10 million Rec-1 cells. 180 days after start of treatment 100 % of the mice treated with chHH1 were still alive, while 70 % of the mice treated with HH1 were alive. All control mice were dead after 80 days. The difference between the chHH1 and the HH1 treatment was not statistically significant. The observed difference in efficacy is likely related to differences in induction of ADCC and other types of antibody induced immunological toxicity. In a separate study, HH1 was found to bind more strongly than chHH1 to mouse receptor FcgRIIb, whereas for the receptors FcgRI and IV chHH1 bound strongly while HH1 did not bind. This difference in binding to mouse Fc-receptors correlates with the observed differences in efficacy. The chHH1 treatment was equally effective as rituximab treatment (data not shown).

chHH1

ADCC was measured using NK-cells from donor blood. HH1 did not induce antibody dependent cellular cytotoxicity in any of the tested cell lines, as compared with control, while chHH1 induced ADCC in all the tested cell lines, except for DOHH2 cells. The effector cell to target cell ratio was 10:1 and the antibody concentration was 20 mg/ml. CDC was not an active mechanism for any of the antibodies.

Biodistribution in nude mice with Ramos tumor xenografts

Urine

BloodLu

ngsHeart

Liver

Spleen

Kidney

Stomach

Small I

ntestine

Large In

testine

Ovarie

sFe

mur

Muscle

Lymph Node

Injection sit

eTumor

Tiss

ue a

ctiv

ity (%

IA/g

)

0

10

20

30

40

50chHH1-DOTA, CAR: 0.5HH1-DOTA, CAR: 0.5chHH1-DOTA, CAR: 1.7 HH1-DOTA, CAR: 0.9 chHH1-DOTA, CAR: 2.5 HH1-DOTA, CAR: 1.6

Both the murine HH1 and the chHH1 antibodies were conjugated with p-SCN-Bn-DOTA using different DOTA:Ab ratios, which resulted in different CARs. Biodistributions of the ARCs were measured 3 days after injection in mice with Ramos xenografts. All ARCs had a relevant biodistribution in this mouse model, and the retention in blood and uptake in normal organs and tumour were similar. There was a non-significantly higher uptake in tumours for the ARCs with lowest CAR.

Ramos cells were incubated with both ARCs for 4 hours, washed and incubated further for 2 days before the viable cell concentration was measured by flow cytometry. Specific activity was 200 MBq/mg for both ARCs. In the figure the results for ARCs with CARs of 1.7 for the chHH1 and 0.9 for HH1 is shown. The results were similar for ARCs with lower (0.5) or higher (1.6-2.5) CAR.

• The two antibodies had similar affinites (Kd 2.5 nM), internalisation and selectivity to human lymphoid tissues.

• chHH1 bound to all classical human Fc receptors, while HH1 bound only to human FcgRIIa and weakly to IIb.

• chHH1 induced ADCC, while HH1 did not. CDC was not an active mechanism for either of the antibodies.

• chHH1 was a more effective treatment than HH1 in mice with intravenous human mantle cell lymphoma xenografts.

• The uptake in normal organs and tumour xenografts were similar for both antibodies.

• The cytotoxicity in vitro was similar for both antibodies.

• chHH1 was predicted to be less immunogenic than HH1.

Control 1 2,5 5

Viab

le ce

ll con

cent

ratio

n rela

tive t

o con

trol c

ells

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

1,6

ControlchHH1, CAR: 1.7HH1, CAR: 0.9

Control HH1 chHH1

% sp

ecifi

c lys

is

0

10

20

30

40

50

60

70Rec-1 DOHH2 Ramos Daudi

anti-CD37 antibody

Lutetium-177

radiation

CD37 antigen CD20 antigen

Tumour Cellnon-Hodgkin Lymphoma

Betalutin® – Mechanism of action

The binding of chHH1 and HH1 to Ramos cells was similar. The chelator to antibody ratio (CAR) was 0.5 in this experiment. The results did not change significantly with higher CARs. The cells were incubated with the ARCs for 1 hour on ice before washing and measurement of specific binding. The binding affinity of both antibodies was 2.5 nM.

HH1

Antibody concentration ( g/ml)0 1 2 3 4 5

OD62

0 nm

0,0

0,5

1,0

1,5

2,0

2,5

3,0

Fc RI Fc RIIa-H Fc RIIa-R Fc RIIb Fc RIIIa-F Fc RIIIa-V Fc RIIIb

chHH1

Antibody concentration ( g/ml)

0 1 2 3 4 5

OD 62

0nm

Antibody concentration (mg/ml) Antibody concentration (mg/ml)

Binding to Fc-receptors was measured using ELISA and recombinant Fc-receptors. The chimeric antibody bound to all the classical human Fc-receptors, while the murine antibody only bound to FcgRIIa and IIb, which are expressed on B-cells.

HH1 chHH1

HH1 chHH1 rituximab

The images show localisation of HH1 and chHH1 predominantly inside Ramos cells, while rituximab is localised on the plasma membrane of the Ramos cells. The cells were incubated over night at 37 °C.

DisclosureDahle, J.: Employment, Equity ownership, Patent. Repetto-Llamazares, A.: Employment, Equity ownership. Melhus, K.B.: Employment, Equity ownership. O’Shea, A.: Employment. Generalov, R.: Employment. Andersen, J.T.: Nothing to disclose. Heyerdahl, H.: Employment, Equity ownership.

0

20 000

40 000

60 000

80 000

100 000

120 000

140 000

160 000

180 000

0 5 000 10 000 15 000 20 000

Spec

ific b

indi

ng

(num

ber o

f Ant

iboi

des/

cell)

Antibody concentration (ng/ml)

chHH1, CAR=0.5

HH1, CAR=0.5

Days since start of treatment

Surv

ivin

g fr

actio

n

Antibody concentration (ng/ml)

Spec

ific

bind

ing

(num

ber o

f anti

bodi

es/c

ell)

Viab

le c

ell c

once

ntra

tion

rela

tive

to c

ontr

ol c

ells

Tiss

ue a

ctivi

ty (%

IA/g

)

FcgRIFcgRlla-HFcgRlla-RFcgRllbFcgRllla-FFcgRllla-VFcgFlllb

% S

peci

fic ly

sis

Control HH1 chHH1

Imm

unog

enic

ity sc

ore

Antibody concentration (mg/ml)