The CD30/CD16A innate cell engager AFM13 elicits polyfunctional NK cell

responses effectively triggering memory-like (ML) NK cells against CD30+ targets

Nancy Marin1, Melissa M. Berrien-Elliott1, Michelle Becker-Hapak1, Martin Treder2, Mark Foster1, Carly Neal1, Ethan

McClain1, Sweta Desai1, Julia A. Wagner1, Timothy Schappe1, Lynne Marsala1, Pamela Wong1, Joachim Koch2, Todd A.

Fehniger1.

1Washington University School of Medicine, Saint Louis, MO; 2Affimed GmbH, Heidelberg, Germany

Natural killer (NK) cells are crucial innate immune effector cells that rapidly recognize and eliminate

infected, stressed and malignant cells. One barrier to broadly applying NK cell therapy across many

cancer types is inconsistent cancer cell recognition, which may be overcome by immune cell

engagers. AFM13 is a tetravalent bispecific antibody based on the ROCK® platform characterized by

bispecific tetravalent binding to CD30 and CD16A, with clinical efficacy in CD30+ malignancies.

Additionally, adoptively transferred memory-like (ML) NK cells have demonstrated enhanced anti-

tumor activity that may be receptive to AFM13-based targeting to enhance target cell recognition.

However, our understanding of NK cell phenotype of functional responses triggered via AFM13

remains incomplete. To address these questions, we analyzed healthy donor-derived conventional

(cNK) and ML (IL-12/15/18-induced) NK cell functional responses to CD30+ lymphoma cells treated

with AFM13. Primary cNK cells co-incubated with AFM13-labeled Hut-78 cells demonstrated

increased IFN-g, TNF, and degranulation, compared to Hut-78 cells or Raji (CD30-) targets + AFM13

as a negative control (p<0.01). ML NK cells also displayed enhanced cytokine production and killing

of CD30+ tumor targets when co-incubated with Hut-78+AFM13 (p<0.01). To define the single-cell

specificity of cNK cell responses to AFM13, similar assays were performed using mass cytometry

assessing 39 lineage, maturation, activating and inhibitory receptors, and function-relevant NK cell

markers. We then used CITRUS to define the NK cell subsets associated with increased IFN-γ,

MIP1α, CD107a (SAM, FDR<0.05). CITRUS-based clustering divided the NK cell populations into

two main groups that associated with increased effector responses. Based on back-gating, we

determined that both mature (NKG2A-CD57+) and immature (NKG2A+ CD57-) subsets displayed

enhanced IFN-g (p<0.05), Mip1a (p<0.0001), and CD107a (p<0.0001) in the presence of Hut-

78+AFM13 compared to Hut-78+AFM12 (CD19/CD16A bispecific innate immune cell engager,

negative control), with mature cells producing the most IFN-g and MIP1a compared to immature

AFM13 triggered NK cells (p < 0.05), or NK cells triggered with AFM12 (p < 0.001). AFM13 also

resulted in increased frequencies of IFN-g+CD107a+MIP1α+ multifunctional cells, in both mature and

immature subsets, compared to AFM12 (p<0.05). Finally, within the NK cells stimulated with Hut-

78+AFM13, we observed significantly increased CD57, KIR2DL1, KIR2DL2/2DL3, KIR3DL1

(p<0.01), NKp30, NKp44, and NKp80 (p<0.01) in the IFN-g producing cells compared to IFN-g-

negative cells. Overall, AFM13 enhanced the magnitude and quality of NK cell responses against

lymphoma targets. Collectively, these data indicate that target cell recognition of NK cells can be

significantly enhanced by AFM13, yet influenced by inhibitory receptors expression, maturation state,

and memory-like differentiation.

Results

Abstract

-Natural killer (NK) cells are cytotoxic innate lymphoid cells that display potent effector responses

against tumor cells. However, they are frequently deficient or dysfunctional in cancer patients.

-Strategies to enhance NK cell functionality and tumor targeting are required to fully harness their

anti-tumor potential.

Introduction

Methods

Uns

tim

K562

Hut

-78

Hut

-78+

AFM13 R

aji

Raj

i+AFM

13

0

20

40

60

%

IF

N-g

po

sitiv

e

*

Uns

tim

K562

Hut

-78

Hut

-78+

AFM13 R

aji

Raj

i+AFM

13

0

5

10

15

20

%

TN

F p

osi

tive

*

Uns

tim

K562

Hut

-78

Hut

-78+

AFM13 R

aji

Raj

i+AFM

13

0

20

40

60

% C

D107a p

osi

tive **

Figure 1. AFM13 binding on CD30+ tumors triggers cytokine secretion and degranulation of cNK cells

from normal donors. Frequency of IFN-g, TNF and CD107 positive NK cells induced after stimulation with

AFM13-prelabeled Hut-78 tumor cells. AFM13 triggering does not modify NK cell cytokine secretion or

degranulation when Raji cells were used as targets. Bars represent Mean + SEM. One-way ANOVA. n=3-5

donors. Data of 3 independent experiments. *p<0.05, **p<0.01

Figure 4. AFM13 binding triggers memory like responses with enhanced cytotoxicity against CD30+

tumors. Control and ML NK cells stimulated with AFM13-labeled Hut-78 cells show increased frequency of

IFN-g, TNF and CD107 positive cells compared to control NK cells. AFM13 labeling (closed symbols)

increased specific killing of Hut-78 cells by control but mainly ML NK cells compared to unlabeled (open

symbols) tumor targets. F) Lack of specific killing of Raji cells demonstrated specificity of AFM13-mediated

enhanced NK responses against CD30+ tumor targets. Two-way ANOVA. Bars represent Mean + SEM.

*p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. n=10.

0

20

40

60

80

%

IF

N-g

po

sitiv

e

****

*

********

Unstim

AFM12 AFM13

+ Hut-78

-5

0

5

10

15

20

25

% T

NF

Po

sitiv

e

***

*

Unstim

AFM12 AFM13

+ Hut-78

0

20

40

60

80

100

% M

IP1a

po

sitiv

e **

Unstim

AFM12 AFM13

+ Hut-78

****

0

20

40

60

% C

D107

positiv

e

*****

**

Unstim

AFM12 AFM13

+ Hut-78

tSN

E2

tSNE1

tSN

E2

tSNE1

tSN

E2

tSNE1

tSN

E2

tSNE1

tSN

E2

tSNE1

3:1 1:1 2:10

20

40

60

80

% S

pecific

Lysis

****

****

****

3:1 1:1 2:10

20

40

60

80

% S

pecific

Lysis

****

****

****

****

0

10

20

30

40

50

% IF

N-g

positiv

e ****

****

****

****

Unstim K562

HUT-78

+AFM-130

20

40

60

80

% C

D107a p

ositiv

e

********

Unstim K562

HUT-78

+AFM-130

3

6

9

12

% T

NF

positiv

e

******

Unstim K562

HUT-78

+AFM-13

Conclusions AFM13 significantly enhances NK cell recognition of CD30+ malignancies correlating with

increased NK cell activation

NK cell expression of inhibitory receptors, maturation state and memory-like differentiation

influence AFM13-mediated NK cell response against CD30+ targets

AFM13 pretreatment of tumor targets potentiates ML NK cell effector functions including cytokine

secretion and cytotoxicity

AFM13 pretreatment trigger polyfunctional responses in NK cells compared to AFM12

pretreatment

Mass Cytometry can be successfully applied to evaluate AFM13-triggered functional responses of

conventional and ML NK cells at single cell resolution

The combination of ML NK cells with AFM13 appears to be a promising therapeutic approach for

treating CD30+ malignancies

Figure 2. Functional mass cytometry reveals that AFM13 profoundly alters cNK cell viSNE maps via

activation and cytokine secretion. A) viSNE plot showing the expression of IFN-g, TNF, CD107, MIP1a and

CD16 on CD56+ cells. B) Frequency of IFN-g, TNF, MIP1a and CD107 positive NK cells in unstimulated, and

cNK cells stimulated with Hut-78 cells alone or AFM12 or AFM13-labeled Hut-78 cells. One-way ANOVA. Error

bars represent Mean + SD. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. n=10 donors

AFM13, a first-in-class tetravalent, bispecific

innate immune cell engager is characterized

by tetravalent binding to CD30 and CD16A

(FcgRIIIA).

-AFM13 binding to CD30+ malignancies

potentiate NK cells activation resulting in

enhanced cytotoxicity and cytokine secretion.

NK cells with memory-like (ML) properties

differentiated after a short-term stimulation with

IL-12, IL- 15 and IL-18 display enhanced

functionality and anti-tumor response.

What is the contribution of immune cell engagers along with the enhanced functionality of

ML NK cells in the overall response to tumor targets? What are the mechanisms underlying

this enhanced response?

Enhanced cytokine secretion

Enhanced cytotoxicity

Results

A

B

Figure 3. AFM13 triggering increases cytokine and chemokine secretion as well as degranulation of

both immature and mature cNK cells. A) Representative SPADE analysis showing phenotype of IFN-g

producing cells in response to AFM13 triggering. Mature KIR+ A), CD57+ B) as well as immature

(NKG2A/CD94+) C) cNK cells display robust IFN-g response upon stimulation with AFM13-labeled Hut-78

cells. Numbers next to each node represent the node ID and color indicates median IFN-g expression for each

node. Nodes positive for IFN-g are depicted over their own tSNE map. D) Expression of maturation markers in

the top 2 nodes positive for IFN-g across the donors. E) Pie charts showing cNK cells stimulated with AFM12

or AFM13-labelled Hut-78 cells single or multiple producers of IFN-g, TNF, MIP1a and CD107. Bars represent

Mean + SEM. n=10.

Media

n E

xpre

ssio

n

(Arc

sin

h)

IFNg

TNF

Mip

1a

CD10

7

0.0

0.5

1.0

1.5

KIR

2DL1

/DS1

KIR

2DL2/

2DL3

CD57

NKG2A

CD94

0.0

0.7

1.4

2.1

2.8

3.5

IFN-ghigh nodes

This study was partially supported with research funds provided to Washington University by Affimed

1+

2+

3+

4+

AFM12(Total=34.2%)

AFM13(Total=58.5%)

D E