ESMO IO 2021 Poster 58P: Achilles VELOS Process 2 generates a >10-fold improvement in cNeT dose over Process 1 with a highly potent polyclonal phenotype

Achilles VELOS Process 2 generates a >10-fold improvement in cNeT dose over Process 1 with a highly potent polyclonal	58P
phenotype and has been successfully validated at GMP scale for clinical use in solid cancer

Evi Rologi1, Monica Sassi1, Samuel Jide-Banwo1, Joseph Robinson1, Michal Pruchniak1, Henrieta Fraser1, Eleni Kotsiou1, Sarah Thirkell1, Malvin Siew1, Sameer Mistry1, Janke Pieters1, Luke Goodsell1, Connor Carolan1, Matilde Saggese1, Michael Grant1, Bethany Samways1, Pooja Kotecha1, Andreas Schmitt2, Farah Islam4, David Lawrence4, Martin Forster3,4, Samra Turajlic2, Mark Lowdell3, Karl S. Peggs1, Sergio A. Quezada1, Edward Samuel1,5

1. Achilles Therapeutics UK Limited, London, United Kingdom; 2) Royal Marsden NHS Foundation Trust, London, United Kingdom; 3) University College London Cancer Institute, London, United Kingdom; 4) UCLH NHS Foundation Trust, London, United Kingdom; 5) Corresponding author - for further information please email: [e-mail]

Introduction

Results

Adoptive transfer of ex-vivo expanded Tumour-Infiltrating Lymphocytes (TIL) has shown promise in delivering durable responses in several solid tumour indications.

Figure 2: VELOSTM Process 2 generates higher TIL numbers following culture of tumour fragments

Figure 4: VELOSTM Process 2 generates a high purity CD3+ T cell product that retains both CD4+ and CD8+ T cell subsets

Figure 6: VELOSTM Process 2 cNeT reactivity and polyclonal stimulation was comparable to VELOSTM Process 1

However, characterisation of the active component of TIL therapy remains challenging due to the non-specific expansion of TILs in the manufacturing process. Achilles Therapeutics has developed ATL001, a patient-specificTIL-based product, manufactured using our VELOSTM manufacturing platform (Figure 1) that specifically targets clonal neoantigens, a subset of patient specific mutations present on all tumour cells. Process 1 of the VELOSTM manufacturing platform has successfully demonstrated the feasibility of generating clonal neoantigen reactive T cells (cNeT) products for the treatment of advanced NSCLC (CHIRON, NCT04032847) and recurrent or metastatic melanoma (THETIS, NCT03997474). Here we report on the successful GMP validation of VELOSTM Process 2 and demonstrate the generation of a significant dose boost of highly potent and reactive CD8+ and CD4+ cNeT for future clinical use with a comparable vein-to-vein time to Process 1.

TIL yield per gram tumour

1010✱

109

108

107

106 Process 1 Process 2

Generation of pre-REP TIL intermediates at GMP scale with VELOSTM Process 2 generated a 10-fold increase in TIL yield compared to VELOSTM Process 1 (Process

1. 18.2 x 106 TIL per gram vs. Process 2: 187.7 x 106 TIL per gram). Lines at median;

*p<0.05 two-tailed Mann Whitney test.

VELOSTM Process 1: n=8,

VELOSTM Process 2: n=10,

(%)	A				NK (CD3-CD56+)	B
				CD19+
cells				CD11c+
100				CD3+				57% CD4+
										43% CD8+
+
								Process 1
CD45
50
of											30% CD4+
Proportion		Process 1 Process 2
				70% CD8+

Process 2

0

Identity and purity (A) were assessed in both infused products manufactured with Process 1 (n=8) and GMP validation runs with Process 2 (n=4). CD3+ T cell purity was comparable in both processes (97.8% vs. 98.6%) with low levelsv (<2%) of cell impurities. Final

		A		B				ns
		100							Producing		100
Producing	(%)Cells			ns	(%)Cells
		80									80
CytokineTotal	CD3	60							CytokineTotal	CD3	60
	T	40								T	40
	+								+
		20									20
		0									0	Process 1 Process 2
		Process 1 Process 2

Figure 7: VELOSTM Process 2 express similar levels of immune checkpoint molecules to initial process development runs

Methods

Figure 1: VELOSTM Process	•	GMP Process 2 data was
		generated from tumour	and
		blood samples procured at the
		time of surgery through our
		ethically approved non-clinical
		study	(NCT03517917)	in
		addition to excess material from
		clinical samples in NSCLC (n=8)
		or metastatic melanoma	(n=2)
		patients.

• Process 1 GMP data was generated from manufactured products in eight patients dosed to date across THETIS (n=5) and CHIRON (n=3).
• TIL were isolated from tumour fragments in the presence of IL-2 and Dendritic Cells (DCs) generated from whole blood.
• Patient-specificpeptide pools were synthesised, corresponding to the clonal neoantigen mutations identified using our proprietary PELEUSTM bioinformatics platform.
• cNeT were expanded by co-culture of TIL with peptide-loaded DCs.
• For VELOSTM Process 2 additional media supplements were added throughout the process. Cell expansion was boosted at the end of the co-culture with an optimized stimulation cocktail.
• cNeT reactivity was assessed using our proprietary potency assay with peptide pool rechallenge followed by intracellular cytokine staining. Product characterisation and phenotype was assessed by flow cytometric analysis.

Figure 3: VELOSTM Process 2 delivers a >10-fold increase in cNeT doses compared to VELOSTM Process 1

		A		B
cellTCD3+ expansionFold									cells)(reactivedosecNeT	109				✱
1000				✱✱
										108
													1
	100									107
	10									106
										105
	1	Process 1 Process 2		104
			Process 1	Process 2

TIL intermediates were used for the selective expansion step with peptide-loaded DCs. Process 2 resulted in a 55-fold increase in the total number of CD3+ T cells compared with a 7-fold increase in Process 1 (A; lines at median, ** p<0.005 two-tailed Mann Whitney test). The active drug component was quantified using Achilles' proprietary potency assay and showed a >10-fold increase with Process 2 yielding a median in cNeT dose of 203.5 x 106 compared with 14.2 x 106 in Process 1 (B; lines at median, * p<0.05 two-tailed Mann Whitney test).

For both figures: VELOSTM Process 1: n=8, VELOSTM Process 2: n=4, 1Extrapolated to reflect minimum tumour volume for clinical procurement

products contained a mix of both CD4+ and CD8+ T cells with a trend to more CD8+ T cells with VELOSTM Process 2 (B).

Figure 5: VELOSTM Process 2 generates a polyfunctional cNeT product at GMP scale

ICS Peptide rechallenge CD8+ T cells

ICS Peptide rechallenge CD4+ T cells

cNeT reactivity in CD4+ and CD8+ T cells was quantified by cytokine secretion in response to peptide rechallenge to determine the active component and further characterisation through measurement of CD107a (Figure 5). Comparison of cNeT reactivity (Figure 6 A) and T polyclonal stimulation with SEB (Figure 6 B) was comparable between VELOSTM Process 1 and Process 2. Figure 6 A and B, lines at median, no significance different, two-tailed Mann Whitney test).

VELOSTM 1 PD

100 VELOSTM 2 PD

of cNeT (%)	VELOSTM 2 GMP
50
Proportion
	0
	CD366+	CD279+	CD278+	TIGIT+

Restimulation with clonal neoantigen pareptide pools and staining for cytokine secreting cells enables further characterisation of the active cNeT drug component (Figure 7). Immune checkpoint molecules were comparable between products manufactured with VELOSTM Process 2 at GMP scale (n=3) and small scale non-clinical process development batches generated with VELOSTM Process 1 and Process 2. Bars at median.

Conclusions

• We performed four VELOSTM Process 2 runs at scale to GMP standards and all batches successfully met all QC release criteria.
• This data demonstrates the ability to generate significantly higher cNeT doses at GMP scale using VELOSTM Process 2 and accurately identify the active drug component.
• Successful GMP Validation of Process 2 supports transfer into a clinical setting for the treatment of advanced NSCLC and melanoma in two first-in-human studies with potential utility across a variety of solid tumours.

© Achilles Therapeutics UK Ltd. 2021