Traws Pharma : Targeting the RAS/RAF/PI3K pathway for CD40 induction in melanoma cells to overcome resistance to anti-PD1 immunotherapies

Targeting the RAS/RAF/PI3K pathway for CD40 induction in melanoma cells to overcome

resistance to anti-PD1 immunotherapies

Chi Yan1,2*, Weifeng Luo1,2, Jinming Yang1,2, Sheau-Chiann Chen3, Bergdorf Kensey1,2, Douglas B. Johnson4, Gregory D. Ayers3, Ann Richmond1,2

1Tennessee Valley Healthcare System, Department of Veterans Affairs, Nashville, TN; 2Vanderbilt University School of Medicine, Department of Pharmacology, Nashville, TN;

3Division of Cancer Biostatistics, Department of Biostatistics, Vanderbilt University Center for Quantitative Sciences, Nashville, TN;

4Division of Hematology and Oncology, Department of Medicine, Vanderbilt University Medical Center, Nashville, TN. * Correspondence: [e-mail] (C.Y.)

Abstract

The first-line treatment for metastatic melanoma is immune checkpoint blockade (ICB), but it fails in many patients and may have serious adverse events. We showed that rigosertib (RGS), a RAS-pathway

RAS/RAF/PI3K inhibition triggers CD40 induction in melanoma cells

CD40 transcription

Tumor Microenvironment

CD40 overexpression in melanoma cells shifted αPD1-non-responding melanomas into responders

CRISPR/Cas9 CD40 lentiviral activation system	Melanoma-cell-specific
(Santa Cruz Biotechnology)	CD40-overexpression(CD40-OE)	Isotype ctrl	Isotype ctrl
MFI=66.2	MFI=94.6

inhibitor, promotes CD40 upregulation on melanoma cells and synergizes with ICB in preclinical melanoma models. However, residual disease still occurred with uncontrolled ERK activation. Our present study explores the efficacy of RGS plus trametinib (T), a MEK1/2 inhibitor, to overcome ICB resistance. The concurrent RGS+T+αPD1 resulted in an additive effect to suppress ICB-resistant NRASmut 1014 tumor growth, and 44.4% (4/9) of the tumors completely regressed after 2 weeks of treatment. The RGS+T regimen promoted CD8+ T cell responses in the tumor microenvironment (TME) and extended the time to αPD1 resistance (p<0.0001). Combining RGS+T with agonist CD40 (aCD40) resulted in increased CD8+ T cells,

	Human BRAFmut Hs294T
activity	800000	melanoma cells
700000		*
		*p<0.05 refers to	*
	600000	*
promoter		RGR+T vs either
500000	RGS or T
CD40	400000
300000
	200000

Drug concentration (nM)

NFκB promoter activity

Human BRAFmut Hs294T

5.00E+06	melanoma cells
			*
4.50E+06	*p<0.05 refers to *
4.00E+06
RGS+T vs T,
3.50E+06	but not RGS

3.00E+06
2.50E+06
2.00E+06
1.50E+06	T
1.00E+06	RGS
5.00E+05	RGS+T
0.00E+00

Drug concentration (nM)

1014

Parental

clones

Control

MFI=93.9

Parental

Parental

CD40-OE clones

MFI=437

clones

Viral ctrl

1

2

3

5

7

11

4

8 9

10 2 3

Viral ctrl

MFI=94.3

MFI=378

37KD

CD40-OE #3

CD40-OE #3

CD40

MFI=1380

MFI=420

50KD

CD40

46.8-fold

PD-L1

37KD

β-actin

increase

Synergistic Activation

Isotype ctrl

Isotype ctrl

Mediator (SAM)

MFI=45.7

MFI=302

B16F10

Parental

Transcription

clones

Activation System

CD40-OE clones

Control clones

MFI=87.5

Parental

Parental

12

9

8

1

14

1

22

23

4

MFI=340

Melanoma

MFI=87.4

Viral ctrl

Viral ctrl

37KD

CD40

MFI=432

cells

CD40-OE #14

CD40

CD40-OE #14

(NRASwt

50KD

MFI=5549

MFI=403

37KD

β-actin

132.0-fold

& NRASmut)

CD40

PD-L1

increase

In vivo tumor growth and therapy responses RNA-Seq at baseline (Veh+IgG): CD40-OE vs CD40-Ctrl

Hallmark_Oxidative Phosphorylation

Hallmark_Myc Targets V1

Hallmark_Hypoxia

Hallmark_E2F Targets

Hallmark_Glycolysis

Hallmark_DNA Repair

natural killer cells, and M1 macrophages in the TME, with a reduction of myeloid-likeCD11b+PD-L1+ regulatory B cells in the tumors (~70%, p<0.0001) and tumor-draining lymph nodes (~40%, p<0.0001). CRISPR/dCas9-based overexpression of CD40 (CD40-OE) in 1014 melanoma cells reduced in vivo tumor growth and successfully turned the ICB-resistant tumors into responders to αPD1 (p=0.025), which further regressed with aCD40+αPD1 (p<0.001). Our preclinical data support the therapeutic use of RAS/RAF/PI3K inhibition plus CD40 agonism for metastatic melanoma patients who do not respond to ICB.

CD40 protein at cell surface level (B16F10 cells)

CD40 MFI

+

B16F10 cell viability

CD40 %

10

****

RGS+T

RGS + Trametinib

%

8

***

T

**

Death

6

Trametinib

***

RGS

*

RGS

*

4

2

Veh

H

O

2

0

-

+

200

400

600

800

0

20

40

60

D

0

D40

4

(Yan et al.,CD40Mol Cancer-MFI of2021)live cells

CD40

+

% of live cells

C

C

RAS/RAF/PI3K inhibition overturns αPD1-resistance in NRASmut melanoma

		1014 NRAS	mut	Tumor	Experimental Design				In combination with αPD1
)	4000	Isotype IgG (n = 7)						)		Veh		T		RGS	RGS + T
3	ICB-resistant 1014 NRASmut melanoma model (n=10)		3	3000		3000		3000		3000
					)	)	)
(mmevolum Tumor		αPD1 (n = 6)				Tumor volume(mm	2500	2500	2500	2500
								3	3	3
3000	αPD1 + αCTLA4							2000	Volume (mm	2000	Volume (mm	2000	Volume (mm	2000
				Vehicle (Veh)
	(n = 7)					***		1500	1500	1500	1500
2000				αPD1	αPD1 +	Trametinib (T)		1000	1000	1000	1000
					500	500	500	500
1000				(200μg)	(200μg)	Rigosertib (RGS)	***	adj.p=	200	200	200	200
						0.052	100	100	100	100
								0	0	0	0

Hallmark_Fatty Acid Metabolism

Hallmark_mTORC1 Signaling

Hallmark_Myc Targets V2

Hallmark_Unfolded Protein Response

Hallmark_Complement

Hallmark_Apical Surface

Hallmark_Mitotic Spindle

Hallmark_IL2/STAT5 Signaling

Hallmark_Apical Junction

Hallmark_IL6/JAK/STAT3 Signaling

Hallmark_Inflammatory Response

Hallmark_Allograft Rejection

Hallmark_IFNα Response

Hallmark_IFNγ Response

-202

Treatment effect on the reduction of

TCRβ-Seq

tumor growth rate compared to Veh+IgG

Distribution of clonotype abundances

Diversity estimation using Chao1

slopeEstimatedchange growthtumorof curve

clonotypesofNumber

IgG αPD1

Diversityof Clonotypes

CD40-Ctrl

***

CD40-Ctrl

CD40-OE

CD40-OE

adj.p=0.07

1

2

3 1

2

3 1

2

3 1

2

3

Abundance

IgG

αPD1

Regulatory B cells, αPD1-resistance and PDO model

0	RGS + T	5	1	2	6	9	3	7	0	5	1	2	6	9	13 17 20	5	1	2	6	9	3	7	0	5	1	2	6	9	13 17	20
	- -				1	1	2	- -				- -				1	1	2	- -

TME immune profiling by FACS

scRNA-Seq of human melanoma

1	4	8	11 15 18 21	Day -12	-5-1 Week 1	Week 5	Day 20	Days ofDaysTreatmentof TreatmentDays of Treatment
	Days of Treatment				RegressionComplete = 0/8	1/8	0/7	4/9

CD8	+	Tc cells

15	**	*

Gated on CD19+ B cells in 1014 melanoma tumors

Human Melanoma scRNA-Seq Dataset

	(Sade-Feldman M., et al, Cell 2018)
cells	pre-treatment samples, n=17

RGS synergizes with ICB in vivo

αPD1-resistance-free probability reaches 50% at

Day 6 13 13 27 (p<0.001)

In tumors

CD8

+

+

Tc cells

CD69

+

Tc cells

CD206

+

+

M2Φ

CD3

)Φ M(

F4/80

60

*

*

150

***

100

*

***

**

*** ***

**

CellsT of%

CellsTc of%

*

Macrophages

80

***

40

100

60

20

50

40

20

0

0

of

0

In TDLNs

h

T

S

T

eh

T

S

T

%

h

T

S

T

e

RG

+

G

S+

e

RG

+

V

S

V

S

V

+

+

+

R

+

CD11c

DCs

+

+

MHCII

CD4 CD3 Th cells

MHCII F4/80

G

G

G

M1Φ

1.0

R

R

R

Cells

*

60

*

30

*

0.8

***

**

CellsT

55

of MΦ

***

**

+

50

20

CD45

0.6

45

RGS + T + αPD1 is

well tolerated in mice

In combination with αPD1

	Mouse Body Weight
25	Veh	RGS
	T	RGS + T
Gram
20

15

ellsc

10

**

+

**

CD45

**

*

5

of%

0

gG gG

0

40

D1

+

1

1

1

G

G

D

D

D

g

Ig

I

I

CD11b

I

B cells

+

+

D4

D

P

P

P

+

+

C

+

+

+

h

h

+aC

a

h

h

0+P S

S

e

e

e

e

40 4

V

V

eh

V

V

*

RG RG

l

E

CD

r

l

l

eh+

E

*

E

t

10

V

V

r

r

t

t

C

-O l

C

O

+a +aCD

-O

***

-

r

-

E

C

-

0

h

h

-

0

40

40

Ct -O

40

4

*

e

0

e

ells-

80

V

4

D4

*

CD

D

0

CD

*D

V D*

C

C

4

rl

E

C

C

CD

C t

*

+

C

-O

CD4

-

40

6

0

CD45of

D4

CD

4

C

2

%

0

PD-L1

MFI % 82.0 92.4 90.8 89.5 80.4 0

PD-L1

		Correlation between
		+	B cells and CD8	+	Tc cells
	CD11b
		8	r=-0.589
cellsB	cells	6	p<0.0001
+	+
D11b%C	CD45in	4
		2

B	100
cellsB	80
+
L1PD
60
+	totalin	40
CD11b
20
0
of
	1	2	3 4	6	7	8	0	4	6	8	9	1	33	35
%		P	P P P P	P	P		2	2 2	2	3
							P12P15P2 P	P P	P P	P P

Patient ID

Mean expression of CD8a in the tumor	r = -0.555				αPD1
cells		non-responder
p = 0.000
		*
		120
	B	cellsB talto		paired t-test
	+	100
	PDL1
	80
	+	60
	D11b

0.4

f

o%

40

%

10

of

0.2

35

%

0.0

30

0

h

T

GS

S+T

h

T

S

T

αPD1 +

αPD1 +

h

T

S

T

αPD1 +

Ve

e

G

+

V

S

e

G

+

V

R

R

G

G

R

RGS

R

R

-5	2	6	9	13 17 20
	Days of Treatment

+

D

0

0

1

1

1

+

G

D

P

PD1P

P

+

G

0

4

4

D

D

D

G

gG g

g

g

I

I

h+ h+

I

I

h+ h

a

a

S

S

e

C

e

e

40+40+

0

5

10

15

V

Ve

C

G

G

+

+

r

eh+eh+

V

V

CD C

E

rl

R

%CD8

Tc cells in CD45

cells

t

r

l

D

R

-

l

O

-O

V

V

t

-OE

0

C

h+ah+a

Ct

E

0

trl

-

0

4

4

C

-OE

C

e

e

-

4

D

40

D

D

-

0

D40 D l

V

D

C

C

r

V

E

40

C

40

4

C

C

C

D

D

Ct

-O

C

C

-

0

0

4

4

D

D

C

C

Patient-derived organoid 3D cultures

	C in	40
	20
CD11b+PDL1+ B cells%	of
in the tumor	%	Pre Post

ICB, αPD1+αCTLA4

(Yan et al., Mol Cancer 2021)

1014 NRASmut
melanoma		****				*
	4	Veh		0.08
pERK/ERK	3	****		T			***
	pAKT/AKT
		RGS	0.06		**
2	*	*	RGS+T		*
	0.04	*

Combine agonist CD40 (aCD40) and RAS/RAF/PI3K inhibition to overcome αPD1-resistance in NRASmut melanoma

Experimental Design

22

Mouse Body Weight

Veh+aCD40

Treatment starts

20

RGS+aCD40

T+aCD40

Veh

Gram

RGS+T+aCD40

αPD1

low dose

RGS

18

aCD40

+

T

(200μg)

(30μg)

16

RGS + T

Transient toxicity of aCD40

Day -12

-5-1

Week 1

Week 4

14

11

14

18

22

25

8

Days of treatment

Toxicity (endpoint)

Tumors

TDLNs

(Vilgelm et al., iScience 2020)

	Patient 2624, primary tumor		Patient 3453A, LN		Patient 3529, primary tumor			Patient 3529, LN			Patient 3101, LN
			***	Brafmut			Brafmut				WT			p = 0.1114	WT			History of renal transplant and	WT
	120000				120000			300000		*		200000			60000
)			)		)		)	p = 0.1149		)	chronic immunosuppression
		***	*		*	250000	**
2				2		2		2				2
(um	100000		*	(um	100000		(um	200000			(um		p = 0.1001		(um	50000
						150000
			**		**
aAre	80000				Area	80000	Area	150000			Area				aAre	40000
60000	**				60000	*	100000			100000			30000
Organoid			***	Organoid	Organoid				Organoid			Organoid
20000				20000		20000						10000
	40000			***		40000			60000								20000
						40000				50000

0

0

0

0

0

l

0

1

0

1

D1

trlGS 40

0

l S

0

1

40

1

D1

l S

0

1

0

1

1

S 0 1

0

1 1

tr GS 4

PD

4

4

D1

tr G

4

PD

4

D

4

D

D

trlG

D

4

PDPD

C

a

+

PDP

C

aCD

PD1D

P

PD1

C

aCD

PDP

Ctr G D

a

D

+

C

a

D4 P

D

a

α C

+

+

α

aCD+

+

aC

+

a

+

+

R

CDα

CDα α

R

α

α

R

α

α

R

α

C

α

α

R

C α

C

α

α

+ S

+a

0

GS+

S 0

P

P

P

+

0

G

G

+ S 0

D

R

D40+

RGS

D4

R

4

R

R

R

S G

R

S

CD4

R

G

S

G

S

GS 4

aC

C

RG

CD

RG

C

a

a

a

a

S+

+

+

S+

S+

S

S

RG

G

G

G

G

R

R

R

R

1					0.02
0		60min	24h	0.00
		60min 24h
			n	4h
		i
5		m		2
0
1
6

RGS-inducedanti-tumor immunity depends on T cells and CD40 in melanoma cells

YUMM3.3 BRAFmut melanoma
	Individual Tumor Volume	Tumor growth in nude mice		CD40 knockdown in
1500	Vehicle (H2O)	)1000	Athymic mice_Veh	)	melanoma cells
	300 mg/kg RGS	3		800

600

AST

U/L

400

200

0

40

0

0

0

4

4

4

D

D

D

D

Tumors+ +

C

+

C

C

C

a

a

a

a

V

h

GS

T+

T

+

e

R

GS

R

+

CD45

cells

Cells

5

*

4

Live

3

2

of

1

%

0

aCD40+

h

T

S

T

Ve

RG

GS+

R

100

ALT

40

BUN

U/L

80

mg/dL

35

60

30

40

25

20

20

0

15

40

0

0

0

40

0

D40

4

4

4

4

D

D

CD

D

+aCD

D40 D

C

C

+

C

C

aC

+

C

a

a

+

a

a

a

a

h+InS+combinationT Th withS+ T+aCD40T

Ve

S+

V

G

+

RG

G

All

e

R

GS

R

Th

R

Tc

B

NK

cells

M2

DC

tSNE2

M1

Gran

Mono

tSNE1

Tc cells

CD69+ Tc cells

M1 Macrophages

Tc cells

CD69+ Tc cells

M1 Macrophages

Cellsof%T

60

***

CellsTcof%

50

*

Macrophagesof

30

*

**

Cellsof%T

45

CellsTcof%

4.0

***

Macrophagesof

40

*

*

10

***

2.0

*

10

40

30

20

40

3.0

*

20

20

2.5

20

10

35

0

0

%

0

30

1.5

%

0

R

S+T

R

S

RGS S+

R

Ve

R

Veh

RG

S+T

Veh

V

V

h

Ve

T

T

GS

h

T

S

+T

T

T

h

T

S

T

h

T

S

+T

S

e

G

e

G

G

GS+

G

G

RG

RG

GS

R

R

R

R

Veh

T

RGS

RGS + T

CD45+ cell

B cells

B

B

B

B

CD8+ Tc cell

llsCe

25

***

CD4+ Th cell

cells

cells

cells

cells

NK cell

20

**

B cell

+

15

Gran

CD45

10

Myeloid cell (other)

Mono

of

5

M1 macrophage

%

0

M2 macrophage

h

T

T

DC

aCD40+

V

RGS

+

e

n=25,000 CD45+ cells per treatment group

RGS

Conclusions and future works

• Inhibition of RAS/RAF/PI3K pathway triggers CD40 upregulation in melanoma cells.
• Melanoma-cell-selectiveinduction of CD40 successfully shifted αPD1-non-respondingNRASmut melanomas into responders.
• Combining RAS/RAF/PI3K targeted inhibition and agonist CD40 offers considerable promise for ICB-resistantNRASmut and NRASWT melanoma treatment.
• In collobration with OncoNova and Merck, we have initiated a Phase IIb clinical trial at Vanderbilt of rigosertib plus pembrolizumab for treatment of patients with unresectable/ metastatic melanoma that is ICB-refractory. Correlative lab studies are ongoing.
  (ClinicalTrials.gov Identifier: NCT05764395, RECRUITING)

										Tumor volume (mm		Athymic mice_RGS			3		shRNA_Ctrl Veh
										800			Tumor volume (mm
										WT mice_Veh					shRNA_Ctrl RGS
3	1000										WT mice_RGS			600	CD40 shRNA1 Veh
								600		****	CD40 shRNA1 RGS
															CD40 shRNA2 Veh
															400	CD40 shRNA2 RGS
						CR: 1/15	400
	500
mm									200					200			*
								0					0
	0
	0 5 8 2 5 7	0	5	8	2	5	7		1	5	8	12	15			1	5	8	11 14 17
	1	1	1				1	1	1		Days of Treatment			Days of Treatment
	Days of Treatment

(Yan et al., Mol Cancer 2021)

TDLNs

In combination with aCD40

CD45+ cells

All

Veh

T

RGS

RGS + T

B cells (53.1%)

B cells (51.0%)

B cells (40.2%)

B cells (42.6%)

CD45+ cell

B cells

100

CD8+ Tc cell

Cells

50

*

NK cell

95

40

CellsLiveof%

CD4+ Th cell

CD45of%

Tc (12.1%)

Tc (15.4%)

Tc (16.7%)

Tc (14.6%)

Macrophage

0

90

B cell

+

30

85

Myeloid cell (other)

20

Gran

80

tSNE2

Mono

10

+

75

T

DC

aCD40+

Veh

T

GS

+T

S

T

aCD40+

Veh

RG

R

RGS

S

RG

tSNE1

n=100,000 CD45+ cells per treatment group

Acknowledgements

• This work was supported by grants from the Lloyd Foundation Melanoma Research Grant (CY), NCI R01-CA116021 (CY, AR), the Department of Veterans Affairs SRCS Award IK6BX005225 (AR) and the MERIT Award 101BX002301 (AR), as well as the VICC Cancer Center Support Grant, P30 CA068485 (AR), for Core Facilities.
• Rigosertib was kindly provided by Onconova Therapeutics, Newtown, PA 18940.