Exploring anti-PD-1 resistance mechanisms for therapeutic t…

Exploring anti-PD-1 resistance mechanisms for therapeutic targeting in NSCLC Haoyi Wu 1 , Jessica M. Konen 1,2 , B. Leticia Rodriguez 1 , Yanhua Tian 1 , Susan Tsang 1 , Jared J. Fradette 1 , Laura Gibson 1 , Don L. Gibbons 1 1. Department of Thoracic/Head & Neck Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX 77030 2. Department of Hematology & Medical Oncology, Emory University, Atlanta, GA 30322

Abstract

Results

Results

A

B

A

B

Immune checkpoint blockade is a promising treatment option for patients with non-small cell lung cancer (NSCLC), the most common type of lung cancer. Lung tumors harboring Kras/p53 (KP) mutations express higher levels of PD-L1 and respond better to anti-PD-1/-PD- L1 therapy than other Kras subsets 1 . However, despite the success of immunotherapy in treating late-stage lung cancer, tumors can gain acquired resistance through mechanisms that are not well understood. Therefore, our goal was to elucidate the mechanisms of intrinsic and acquired anti-PD-(L)1 resistance in NSCLC. For this purpose, we developed several anti-PD-(L)1 sensitive and resistant mouse cell lines as working models. Preliminary studies suggest that our resistant cell lines do not exhibit known mechanisms of tumor cell-intrinsic immunotherapy resistance. To further identify genes that drive novel resistance mechanisms on a single cell resolution, we studied the differentially expressed genes (DEGs) between 344SQ (sensitive) and 344SQ PD1R1 (resistant) tumors treated in vivo with either IgG or anti- PD-1. After validation of numerous DEGs at both the mRNA and protein level, we obtained a list of candidate genes. Interestingly, we found methylthioadenosine phosphorylase (MTAP), a housekeeping gene known to play tumor-suppressor roles, to be consistently and significantly upregulated in anti-PD-1 resistant cell lines and tumors. We hypothesized that gene expression changes in immunotherapy resistant tumor cells reprogram the tumor microenvironment to create an immunosuppressive milieu . Our pilot in vivo study suggests that MTAP knockdown partially re-sensitizes 344SQ PD1R1 tumors to anti- PD-1 treatment and modulates intratumoral T cell activation status. The outcome of this project aims to provide novel therapeutic targets in combination with immunotherapy to overcome anti-PD-(L)1 resistance in NSCLC patients.

T cell phenotype

344SQ PD1R1

344SQ

0 10 20 30 40 50 60 70 80 90 100

PD1R1

344SQ

344SQ

Malignant cells T/NK cells B cells Myeloid cells Neutrophils pDC Fibroblast

0 10 20 30 40 50 60 70 80 90 100

Naïve T cells Memory T cells Exhausted T cells Proliferating T cells T regulatory cells

C

C

D

Tumor growth by treatment

Tumor weight by treatment

Figure 2 . Clustering of single cells using markers from literature identified seven distinct populations. A. Unsupervised clustering of pooled samples identified 7 main cell types. B. Percentage of the identified cell clusters was graphed within each sample. C. Markers from literature used to identify each cluster.

1500

2.0

shCTL IgG shCTL aPD-1 shMTAP IgG shMTAP aPD-1

1.5

1000

1.0

500

0.5

140P IgG vs WT IgG Week 6 DEGs 344SQ PD1R1 vs 344SQ IgG

PD1R1

344SQ vs 344SQ week 6 candidate genes cell line validation

A

B

0

0.0

200 250 300

0

1

2

3

4

8

IgG aPD-1 IgG aPD-1 shCTL shMTAP

Weeks post implantation

Mtap

E

6

150

Naive CD8+

CD8+

Exhausted CD8+

Eff/mem CD8+

4

100

80

50

100

20

80

40

2

50

60

Mtap

15

60

30

-3-2-1 0 1 2 3 0

40

0

10

Background

40

20

0

1

2

3

4

FC

FC (normalized to 344SQ)

20

5

20

10

C

D

MTAP validation in week 6 tumors

344SQ PD1R1 tumor

344SQ tumor

0

0

0

0

A

IgG α PD-1 IgG α PD-1

IgG α PD-1 IgG α PD-1

IgG α PD-1 IgG α PD-1

IgG α PD-1 IgG α PD-1

shCTL shMTAP

shCTL shMTAP

shCTL shMTAP

shCTL shMTAP

Figure 5. Pilot in vivo experiment shows MTAP KD in malignant cells partially re-sensitizes resistant tumors to anti-PD-1 treatment and alters T cell activation state. A. Clustering of T cell population from scRNA sequencing using markers from literature revealed 6 sub-populations with differential stages of T cell activation B. Percentage of each T cell sub-types out of total T cells for each sample. C. Markers used to identify each T cell subcluster phenotype. D. 344SQ PD1R1 shCTL, shMTAP tumor growth curves over 4 weeks (left). Tumor weights at week 4 end point (right). E. Flow cytometry of CD8+ T cells isolated from 344SQ PD1R1 CTL KD and MTAP KD tumors.

344SQ PD1R1 IgG

344SQ IgG

Figure 3 . MTAP is significantly upregulated in resistant tumor in multiple scRNA-seq DEGs comparisons and its expression is validated. A . Volcano plot of week 6 344SQ PD1R1 vs 344SQ IgG DEGs comparison from which MTAP (red) was identified. MTAP was similarly upregulated in week 4 344SQ PD1R1 vs 344SQ IgG and aPD-1 treated DEGs comparisons (not shown). B. A list of candidate genes from T2 comparisons were validated by qPCR in cell lines, with MTAP being the highest upregulated in 344SQ PD1R1 . C. RNA validation of MTAP levels by qPCR in IgG treated week 6 344SQ PD1R1 and 344SQ tumors. D . Protein validation of MTAP levels by IHC in baseline SQ tumors.

T2

T1

T1

T2

B

Conclusions

3. 344SQ + IgG vs. 344SQ PD1R1 + IgG Baseline differences

1. 344SQ + IgG vs. 344SQ + aPD-1

Impact of treatment in parental cell line

• We have developed and validated a working model to study anti-PD-(L)1 resistance in KP mutant lung cancer • scRNA-sequencing analysis identifies high T/NK infiltration in the resistant tumor, but these T cells mostly remain at a naïve state • Anti-PD-1 resistant model has significantly higher MTAP expression and enzymatic activity than the sensitive cell lines • Pilot in vivo study suggests MTAP KD decreases tumor growth at baseline and partially re-sensitizes resistant tumors to anti- PD-1 treatment, potentially by pushing CD8+ T cell activation states closer toward an effector/memory phenotype

B

A

4. 344SQ + aPD-1 vs. 344SQ PD1R1 + aPD-1 Differences between the two cell lines in the face of aPD-1

2. . 344SQ PD1R1 + IgG vs. . 344SQ PD1R1 + aPD-1 Impact of treatment in the aPD-1 res cell line

MTAP cell panel validation

C

Figure 1 . Single cell RNA-sequencing experimental design and validation. A. 344SQ (sensitive) or 344SQ PD1R1 (resistant) tumors were subcutaneously implanted into wildtype mice. One week post implantation, IgG or anti-PD-1 were i.p. injected weekly until end point (week 6). Tumors were collected at Time 1 (T1) (344SQ remained sensitive) and Time 2 (T2) (344SQ gained resistance) to be processed for single cell RNA-sequencing. B. Four comparisons between tumor models and treatments were performed on both T1 and T2 samples to obtain 8 sets of differentially expressed genes (DEGs) comparisons. C. Workflow of DEGs validation, identified from comparisons in Figure 1B.

C

MTAP Specific Activity

***

Future Directions

****

**

1000

D

800

• Optimize the in vivo functional validation experiment with MTAP KD and MTAP OE cell lines to better characterize both the lymphoid and myeloid population in the tumor microenvironment • Identify upstream regulators of MTAP expression in the context of anti-PD-1 resistance • Further in silico analysis using human cancer datasets and in vitro/in vivo functional studies are needed to refine the list of candidate genes

600

400

200

0

344SQ PD1R1 PD1R2 shCTL sh#1 sh#5

PD1R1

344SQ

Figure 4. MTAP expression and enzymatic activity are further validated in different resistance models. A . Validation of MTAP RNA levels by qPCR from panel of 344SQ sensitive and resistant cell lines. B . IHC of MTAP in IgG/aPD-L1 treated tumors derived from sensitive/resistant KP GEMM cell lines. C . Assay based on MTAP (from cell lysate) conversion of MTA to adenine, which was converted to 8-dihydroxyadenine by xanithine oxidase. Absorbance read at 305nm for 30 min at kinetic mode 2 . D . Schematic depicting MTAP housekeeping function and downstream metabolites produced (made with BioRender).

Funding CPRIT RP200235 CPRIT-MIRA RP160562-P3

Contact info E-mail: hwu7@mdanderson.org

References

1. Skoulidis et al. Cancer Discovery . 2015 August ; 5(8): 860 – 877. 2. Christopher et al. Cancer res. 2 002 ; 62 (22), 6639 – 6644.

NIH R37 CA214609 NIH F32 CA239292 2P50CA070907-21A

Page 1

Made with FlippingBook - professional solution for displaying marketing and sales documents online