Transcriptomic Profiling of the Immune Response in Orthotopic Pancreatic Tumours Exposed to Combined Boiling Histotripsy and Oncolytic Reovirus Treatment

Abstract

Background: Boiling histotripsy (BH) uses high-amplitude, short-pulse focused ultrasound to disrupt tissue mechanically. Oncolytic virotherapy using reovirus has shown modest clinical benefit in pancreatic cancer patients. Here, reovirus and BH were used to treat pancreatic tumours, and their effects on the immune transcriptome of these tumours were characterised. Methods: Orthotopic syngeneic murine pancreatic KPC tumours grown in immune-competent subjects, were allocated to control, reovirus, BH and combined BH and reovirus treatment groups. Acoustic cavitation was monitored using a passive broadband cavitation sensor. Treatment effects were assessed histologically with hematoxylin and eosin staining. Single-cell multi-omics combining whole-transcriptome analysis with the expression of surface-expressed immune proteins was used to assess the effects of treatments on tumoural leukocytes. Results: Acoustic cavitation was detected in all subjects exposed to BH, causing cellular disruption in tumours 6 h after treatment. Distinct cell clusters were identified in the pancreatic tumours 24 h post-treatment. These included neutrophils and cytotoxic T cells overexpressing genes associated with an N2-like and an exhaustion phenotype, respectively. Reovirus decreased macrophages, and BH decreased regulatory T cells compared to controls. The combined treatments increased neutrophils and the ratio of various immune cells to Treg. All treatments overexpressed genes associated with an innate immune response, while ultrasound treatments downregulated genes associated with the transporter associated with antigen processing (TAP) complex. Conclusions: Our results show that the combined BH and reovirus treatments maximise the overexpression of genes associated with the innate immune response compared to that seen with each individual treatment, and illustrate the anti-immune phenotype of key immune cells in the pancreatic tumour microenvironment.

1. Introduction

Pancreatic ductal adenocarcinoma (PDAC) presents a formidable treatment challenge due to its exceptionally poor prognosis, low survival, and high disease recurrence rates, leading to approximately half a million deaths worldwide every year [1]. Surgical resection is the current standard of care and is potentially curative, but the proximity of the duodenum and major vessels renders most tumours unresectable. For patients with unresectable PDAC, first-line treatments include the administration of FOLFIRINOX (folinic acid, fluorouracil, irinotecan, and oxaliplatin) [2,3], and, in a metastatic setting, a combination of gemcitabine and nanoparticle albumin-bound paclitaxel is beneficial [4]. However PDAC tumours generally respond poorly to chemotherapy, and in most cases, survival is measured in months rather than years. The application of immunotherapy has yet to demonstrate therapeutic benefits for PDAC patients. This lack of efficacy is due to resistive mechanisms, which include a dense stroma that impedes drug penetration [5,6], and an immune-suppressive tumour microenvironment (TME) [7]. There is therefore a pressing need for adjuvant approaches that could help to overcome the immune resistance of immunologically “cold” PDACs.

Oncolytic viruses (OVs) provide a form of immunotherapy and are known to modulate the TME in tumours stratified as ‘less immunosuppressive’ [8]. These viruses have natural oncolytic properties, being non-pathogenic and replication-competent, and infect cancer cells either through endocytosis or by binding to virus entry surface receptors. OVs trigger a cell’s demise by inducing lysis and immunogenic cell death and are subsequently eliminated by the adaptive immune response [9,10]. The oncolytic Reovirus is being investigated for the treatment of pancreatic cancer, but to date with only limited success. The NCT01280058 phase II ‘Carboplatin and paclitaxel with or without viral therapy in treating patients with recurrent or metastatic pancreatic cancer’ trial [11] investigated the unmodified reovirus type 3 Dearing (Pelareorep) in combination with chemotherapy, with the control arm receiving chemotherapy alone (n = 37). This trial found no significant difference in the progression-free survival between the two groups, although increases were seen in pro-inflammatory immune biomarkers (T and natural killer (NK) cells, cytokines/chemokines) [12] in the combined reovirus and chemotherapy-treated arm relative to the control arm. In another phase II trial (NCT00998322; A study of REOLYSIN^® in combination with gemcitabine in patients with advanced pancreatic adenocarcinoma) 23 out of 34 patients had stable disease following the combined treatment which was well tolerated [13]. A phase I clinical study (NCT03723915; Pembrolizumab and Pelareorep (reovirus) in treating patients with advanced pancreatic cancer) showed significant increases in circulating immune cells and various pro-inflammatory chemokines and cytokines when reovirus was present [14,15]. These results, and the requirement to increase the PDAC infection rates [13], demonstrate the potential for the use of adjuvant therapies, in combination with reovirus, to tap the immense potential for increasing immunogenic PDAC cell death.

Focused ultrasound is a physical modality which can destroy tumours whilst sparing normal tissue [16]. It is a rapidly expanding therapeutic field, and the full “gamut” of its applications can be seen in a report from the Focused Ultrasound Foundation [17]. Focused ultrasound can induce either thermal and/or mechanical effects inside the target tissue and can be switched from predominantly thermal to predominantly mechanical effects by changing the mode of energy delivery [18]. Boiling histotripsy (BH), one of the “mechanical” modes of focused ultrasound tissue damage, induces acoustic cavitation (clouds of highly energetic micron-sized bubbles) which initiates mechanical cellular disruption via bubble collapse [19]. Clinical trials of focused ultrasound-created thermal ablation in PDAC tumours have demonstrated feasibility and safety [20], whereas phase I histotripsy trials for liver, pancreatic and kidney cancers are underway [21,22].

Our group has demonstrated that “mechanical” focused ultrasound using exposures in the BH range [23] (peak negative pressure (P-) ~17 MPa, duty cycle (d.c.) 1%, pulse repetition frequency (prf) 1 Hz, pulse length 10 msec, lesion spacing = 2 mm) in combination with checkpoint inhibitor immunotherapy (anti-CTLA-4 and anti-PD-1 antibodies) can induce anti-cancer immune effects in PDAC tumours by disrupting the tumour core and changing the immune architecture of the tumour to a pro-inflammatory phenotype [24]. Despite the beneficial effects of these treatments, viable tumours remained, and ultimately the subjects succumbed to their disease. Here, we extend our previous study to investigate the acute immune response of murine PDAC tumours exposed to BH treatments using single-cell transcriptomics to elucidate the gene modulation of a variety of immune cell types. To enhance the immune response of these “immune-cold” PDAC tumours, ultrasound treatments were combined with reovirus as an immune adjuvant. We hypothesise that the two treatments, when combined, will show greater immune cell and gene modulation than either of the two treatments alone, and that this will help us better understand the regulation of the various immune cell types and mechanisms associated with the BH treatment response in PDACs.

2. Materials and Methods

Orthotopic PDAC murine model: Murine subjects were purchased from Charles River (UK), and housed in semi-sterile conditions in the animal facilities of the Institute of Cancer Research. Experimental work was carried out under the relevant UK Home Office project and personal licences. Ethics approval was obtained from the local Animal Welfare and Ethics Committee of The Institute of Cancer Research, London. The subjects were monitored regularly for their quality of life (including weight, condition, behaviour, and pain) to ensure the highest welfare standards for the subjects. Their Body Condition Score, the NC3R Grimace Scale, and the modified FELASA Clinical Signs were used to ensure that no animal exceeded the permitted severity limit. Subjects were sacrificed using Schedule 1 humane methods.

The pancreatic KPC cell line (KRASG12D/+; TRP53R172H/+; PDX-1-CRE) of low passage (<5) (kindly provided by Prof. Tuveson (CSHL, NY, USA)) was cultured in Dulbecco’s modified eagle medium supplemented with 2 mM L-glutamine, and 10% foetal bovine serum. All reagents used for cell culture were purchased from Sigma-Aldrich (Gillingham, UK) unless otherwise stated. KPC cells were short tandem repeat (STR) profiled, and routinely tested for mycoplasma. KPCs grown in T75 flasks were harvested using Accutase and 10 × 10⁶ cells/mL were resuspended in Engelbreth–Holm–Swarm murine sarcoma gel. Syngeneic KPC tumours were grown orthotopically by injection of 20 μL of the above cell suspension into the pancreas of C57BL/6 (strain 000064J) mice via a laparotomy (Appendix A). Tumour growth was monitored once a week using the E-Cube 9 (Alpinion Medical Systems, Bothell, WA, USA) 2-D B-mode ultrasound imaging system (central frequency = 14 MHz). PDACs were imaged every 1 mm in the cranial/caudal (sagittal) and medial/lateral (axial) imaging planes. The maximum dimension in 3 orthogonal directions was measured. Typically, tumours could be treated 4 weeks post KPC cell implantation, once all orthogonal dimensions were approximately 7 ± 2 mm.

Experimental studies: For single-cell transcriptomic analysis 24 h post “treatment”, subjects (n = 21) were randomised into four groups—Group 1: sham-exposed, “control” (n = 6); Group 2: reovirus-treated (n = 6); Group 3: BH-exposed (n = 5); Group 4: BH and reovirus “combination” treatments (n = 4). We intended to allocate at least 5 subjects per treatment group; one sample from group 4 (combination) was unfortunately lost during sample processing. For histology experiments 6 and 72 h post “treatment”, 2 subjects were used as controls, and 3 subjects were used for every other group at each time point. For both histology and single-cell transcriptomic experiments, time zero hours is defined as the end of a subject’s “treatment” with BH or sham-BH exposures.

BH exposures (groups 3 and 4) were carried out using the Alpinion VIFU 2000 therapy ultrasound platform (Alpinion Medical Systems, Bothell, WA, USA) (Figure A1). The water tank and subjects were prepared as described in Appendix A. The positions of the PDAC tumour’s upper and lower edges were identified using the 3–12 MHz US imaging probe of the VIFU 2000 system. Subsequently, the central regions of the tumour were located and exposed under ultrasound imaging guidance to avoid damage to adjacent organs/tissues at risk (spleen, stomach, intestine, peritoneum, skin, and abdominal wall). To achieve this, no exposures were placed within ~2 mm of the tumour edges. The number of treated positions per tumour (10 to 30) was dependent on tumour size, and on how well it could be differentiated from its surroundings. Also, to achieve the safe delivery of ultrasound energy in every tumour, every time significant hyperechoic ultrasound signals were seen using real-time ultrasound imaging during exposure, the exposures were immediately stopped, and the treatment was moved to the next planned position. This is because, in our experience, over-treatment of murine orthotopic tumours with BH can have detrimental quality-of-life effects.

The parameters used for BH exposures of each position were 10 msec pulses at a frequency of 1.5 MHz, d.c. 1%, prf 1 Hz, with a spacing between treatment planes and exposure positions in each plane of 1 mm. The free-field P⁻ (~17–19 MPa), electrical power (200–225 W), and number of exposure pulses delivered per position (up to a maximum of 25 and at least 12) varied slightly depending on the bright hyperechoics seen on real-time ultrasound imaging feedback during exposure. Calibration of the VIFU2000 treatment transducer was performed by using a calibrated hydrophone to obtain free-field pressure values as a function of the system’s electrical power settings at low levels (<−5 MPa) to avoid sensor damage and extrapolated to higher levels. Subjects in groups 1 and 2 were sham-exposed by depilating the anaesthetised subjects, and placing them in the VIFU’s heated water bath for 10 min in a similar manner to groups 3 and 4, but without treating them with BH.

For oncolytic reovirus treatments, stocks of T3Dearing strain reovirus were obtained from Oncolytics Biotech Inc, and aliquots of ~3.2 × 10⁹ pfu/mL were banked at −80 °C for long-term storage [25,26]. Mice were incubated at 37 °C for 10 min before reovirus was administered systemically through an IV tail injection in 3.2 × 10⁷ pfu/dose in Hanks balanced salt solution (HBSS), for groups 2 and 4, within 30 min post-BH treatment. Subjects in groups 1 and 3 were sham-exposed using IV tail injections of the vehicle only.

Acoustic cavitation detection and data processing: Acoustic cavitation was monitored in a semi-quantitative manner using a passive cavitation detector (PCD) and a hardware filter which removed the 1.5 MHz drive frequency, as described previously [16,24]. Briefly, the PCD was pulse-echo aligned manually with the BH focus at the beginning of each experimental day (in the absence of the electrical filter). Acoustic emission data were recorded for each pulse, and processed using custom-made MATLAB scripts (version R2020b) which quantified the half-harmonic (0.75 MHz) and broadband signals in the range 0.1 to 3 MHz, following removal of all drive harmonics and subharmonics using a software comb filter. This broadband range was chosen to minimise the detection of imaging signals due to using imaging to detect hyperechoes during exposures. Every PCD detected exposure pulse was divided into 120 cycle segments and fast Fourier transformed to obtain a series of emission spectra. The broadband signal for each exposure pulse was quantified as the sum of voltage amplitudes at each frequency within the broadband range after software (comb filter) removal of harmonic and ultraharmonic signals. Half-harmonic signal for each exposure pulse was the sum of voltages in 3 frequency bins centred on the half-harmonic. Each pulse was characterised using the median amplitude of the half-harmonic and the comb-filtered broadband from all segments. If at least 3 consecutive segment amplitudes were above baseline noise (the average “off-time” signal plus 3 standard deviations) the pulse was identified as a positive (half-harmonic or broadband) event. For each tumour, the total number of exposure pulses and half-harmonic-positive and broadband-positive pulses were calculated. Then the percentage of half-harmonic- and broadband-positive exposure pulses was computed

Histology: For histological analysis, the subjects were culled 6 or 72 h after treatments. The 6 h time point was chosen because the histological effects of BH are likely to be immediate. The 72 h time point was chosen as a reasonable indicator of the delayed tissue effects of BH exposures. The excised KPC tumours were fixed in 10% neutral buffered formalin for 48 h and processed for paraffin embedding overnight. Sections with a thickness of 5 μm were then cut using a microtome, and placed on adhesive slides coated with poly-lysine. These slides were subsequently incubated for 30 min on a 55 °C hot plate, and stored at room temperature. Haematoxylin and eosin (H&E) staining was carried out as described in Appendix A.

Cell preparation and cell sorting for multi-omics experiments: For single-cell experiments, KPC tumours were disintegrated 24 h after treatment using the gentleMACS™ Octo dissociator with Heaters kit and processed for single-cell suspension following the manufacturer’s protocol (Miltenyi Biotech, Bergisch Gladbach, Germany), briefly described in Appendix A. Then, single-cell multi-omics was performed using the BD Rapsody system, combining whole-transcriptome profiling (WTA kit; BD Biosciences, San Jose, CA, USA) with the expression of surface-expressed proteins using the AbSeq technology (BD biosciences). To profile the changes of treatment on the immune compartment, CD45⁺ cells (e.g., T and B lymphocytes, monocytes/macrophages, dendritic cells, and granulocytes, including neutrophils) were isolated and profiled using the BD Rhapsody HT Xpress system (BD Biosciences). Briefly, cell-containing frozen vials were thawed in a water bath at 37 °C and agitated until only a small ice clump remained. X-Vivo medium (1 mL) supplemented with 1% heat-inactivated FBS was then added to each vial in a brisk, drop-by-drop manner. The contents of each vial were transferred into a sterile pre-labelled 15 mL centrifuge tube, made up to 10 mL using additional X-Vivo 15 (Lonza, Basel, Switzerland) medium supplemented with 1% FBS, and mixed by gently swirling, followed by centrifugation at 400× g, for 7 min, at room temperature to pellet the cells. The supernatant was carefully discarded, and 500 mL of ACK lysis buffer was added to the cell pellet and mixed for 1 min. This mixture was made up to 13 mL final volume with ice-cold PBS + 2% FBS and was centrifuged at 400× g, 7 min, 4 °C. The cell pellet was resuspended in 5 mL of cold PBS + 2% FBS. Cells were counted and distributed into 10⁶ cell aliquots. Each of the aliquots was transferred into a sterile pre-labelled FACS tube and topped up with ice-cold PBS + 2% FBS. Tubes were centrifuged at 400× g, 5 min, 4 °C. Pellets obtained were resuspended in 50 mL of BD FACS buffer. Then 2 mL mouse Fc block (1 μg/million cells) was added and incubated on ice for 5 min. Cell aliquots were then labelled with the respective sample barcoding antibody (sample multiplexing kit; BD Biosciences) and with FITC-conjugated anti-mouse CD45 antibody (BioLegend, San Diego, CA, USA) for a further 15 min on ice. 110 μL BD FACS buffer was then added to the tubes (BD biosciences) and cells were stained with DRAQ7 (1:500 dilution) for 15 min on ice for the identification of dead cells. Cells were then washed and resuspended in cold PBS + 1% FBS and filtered for cell sorting at 4 °C in a BD FACSAria Fusion sorter (BD Biosciences). A total of 5 × 10⁴ CD45⁺ DRAQ7⁻ live cells were sorted from each experimental condition. Cells were processed into three experimental days, with 8 samples (duplicates of each of the 4 experimental conditions) being processed on each day.

AbSeq antibody stain and single-cell capture: After cell sorting, barcoded cells from 8 samples (2 samples per experimental condition) were pooled together and incubated with a master mix of oligo-conjugated AbSeq antibodies (BD Bioscience; as listed in

Table A1. List of AbSeq antibodies used in this study showing the antibody target and barcode sequences.

AbSeq Ab	Sequence ID	Barcode Seq
CD172a(SIRPa)	AMM2097	CGTGGTGAGTTGCGAGTGTGCGTATTATTATCTATG
CD19	AMM2007	AAGCATGTCGTTTGTGGCGTACTATTAAGGTGAAGC
CD11c	AMM2008	ATTGGGCGTAAAGGGTAAGGCGGTATATGGACTGTG
CD44	AMM2010	CATGGGTTGTCTCGTTGTAAGTAGTATAGTTGCTGC
TCR B CHN	AMM2021	CAGGTATTAGGAAGATTAGGCCGTTATGATTGGAGC
F4/80	AMM2028	GTCGTGGTCGGATAGCGTGTAGGTTTAAAGTAGAGG
H-2kb	AMM2060	CGGTATATATCTCGGAGGTAAGCGTCGCGGAAATGT
I-Ab	AMM2078	TGAGGTGTTATGTCGTTAGGGTCGTAGTGAAATTGC
Ly-6G	AMM2009	AACAATAGGGATGCGGGATAAGAATACGAAAGGAGT
Ly-6G/C	AMM2015	CATTGCGAGGAGTAAGGCGATATCTAGTTGTGCTGG
CD103	AMM2168	CAATATAATAGCCGGTAGGTGTAGTGCGTAATCAGG
CD3e	AMM2001	GAGATAGGCTAGTTGGATAATTGCGCGGTGAGAGTC
CD4	AMM2002	GTTTAGCGTAGGGTGCATTAGAGCGAGTTAGCGAGT
NK-1.1	AMM2017	GGTCTGGGATTCGTATAGTTCGCGGTAGTTGAGCTT
CD62L	AMM2018	TAGGAGAGATTCGTGGTAGATTTAGCGTAGGTCATT

) for 35 min on ice, according to the manufacturer’s instructions. Cells were then washed three times with BD Sample Buffer at 4 °C to remove residual unbound oligo-conjugated AbSeq antibodies, resuspended in 500 mL cold BD Sample buffer, and filtered for cell counting. Three sample aliquots were then resuspended in 620 mL of cold BD sample buffer at a concentration of 40 cells/mL—for an estimated capture rate of ~15,000 single-cells/cartridge—and immediately loaded on three BD Rhapsody cartridges (BD Biosciences) for single-cell capture.

Transcriptomic data analysis: Fastq files were processed using the Institute of Cancer Research London installation of BD Rhapsody Sequence Analysis Pipeline (v1.11) to generate expression count matrices for RNA and Abseq data, which were analysed in R using the Seurat package (v4.3.0). Quality control was performed separately for data from the three BD Rhapsody cartridges, filtering out cells with fewer than 300 or more than 5000 detected genes, fewer than 500 unique molecular identifier (UMI) counts, or greater than 20% mitochondrial content. Genes detected in fewer than 10 cells were also excluded. After quality control, the filtered datasets were merged for downstream analysis. RNA and Abseq data were normalised separately, with RNA processed using the LogNormalize function and Abseq data normalised using the centred log-ratio (CLR) method. Batch correction and integration of Abseq data were performed using the reciprocal PCA (RPCA) method, followed by feature selection using FindVariableFeatures and data scaling with ScaleData. Dimensionality reduction was performed using PCA on RNA data and significant principal components were selected based on the elbow plot method. To integrate RNA and Abseq modalities, a Weighted Nearest Neighbour (WNN) graph was constructed using FindMultiModalNeighbours. Clustering was performed using the Louvain algorithm (FindClusters) at a resolution of 0.9, to detect distinct immune cell subsets within the TME of pancreatic tumours across all subjects included in the study and clusters visualised using the Uniform Manifold Approximation and Projection (UMAP). FindConservedMarkers was used to identify the top markers for each cluster that were conserved between the four treatment conditions. Cell-type annotations were manually assigned based on established markers and the literature. Pseudobulk differential expression analysis was performed in R using the DESeq2 package, and gene ontology enrichment analysis was conducted with the clusterProfiler R package to evaluate treatment effects on immune cell subpopulations and their associated biological processes.

Statistical analysis: Immune cell abundance and immune cell/Treg ratio data are shown as mean ± standard error of the mean (SEM). The number of experimental replicates is stated in each figure and statistical significance is calculated using a 2-way unpaired equal variance Student’s T test. The statistical significance of differences in transcriptomic data is automatically calculated using p-adjusted values in R (version 4.3.1). In all cases, differences were considered statistically significant at p < 0.05. Unless otherwise stated in the manuscript, gene overexpression was considered “high” if their AvgLog₂Fc > 1, “moderate” if 1 ≥ AvgLog₂Fc ≥ 0.5, “low” if 0.5 > AvgLog₂FC > 0.25, and “negative” if AvgLog₂FC < 0.25.

3. Results

3.1. BH Treatment Monitoring and Acoustic Cavitation Monitoring

BH exposure of tumours resulted in the detection of half-harmonic-positive and broadband-positive signals. In the BH alone group, 17–30 positions/tumour were exposed to BH (on average, 24 ± 6), and 350–720 pulses per tumour were delivered (on average, 550 ± 150). At least 98% of these pulses were positive for half-harmonic signals, and the number of half-harmonic-positive pulses per tumour ranged between 350 and 720 (

Table 1. Acoustic cavitation detection data during and after BH and combination (BH + reovirus) treatments of KPC tumours in every subject. Colour coding has been used to facilitate visual assessment of the numerical data. Each column is colour-coded from lowest value (white) to midpoint (orange) and maximum value (red) automatically in Excel (except for the percentage of HH positive pulses which are all high and therefore coded red).

Subject	Number of Exposure Positions per Tumour	Number of Exposure Pulses per Tumour	Percentage of HH-Positive Exposure Pulses (%)	Number of HH Positive Exposure Pulses per Tumour	Percentage of Broadband Signal Exposure Positive Pulses (%)	Number of Broadband-Positive Exposure Pulses per Tumour	Hyper-Echoic Signals
BH- Subject 1	28	660	98	650	14	95	Yes
BH 2- Subject 2	17	350	100	350	94	330	Yes
BH 3- Subject 3	30	720	100	720	20	150	In-conclusive
BH 4- Subject 4	19	450	100	450	77	345	Yes
BH 5- Subject 5	25	535	100	535	67	360	Yes
Comb 1- Subject 1	13	310	97	300	100	310	Yes
Comb 2- Subject 2	10	230	91	210	95	220	Yes
Comb 3- Subject 3	22	530	100	530	4	20	Yes
Comb 4- Subject 4	17	350	100	350	2	6	Yes

). In the same treatment group, the percentage of BH pulses resulting in broadband signals ranged between 14 and 94% per tumour, and the number of broadband-positive pulses ranged from 95 to 360 per tumour (Table 1). In the combination group, 10–22 positions/tumour were exposed to BH (on average, 15 ± 5), and 230–530 exposure pulses per tumour were delivered (on average, 350 ± 130). Here, more than 91% of the pulses were positive for half-harmonic signals, and the number of half-harmonic-positive pulses per tumour ranged from 210 to 530 pulses. Also, the percentage of broadband-positive pulses ranged from 2 to 100%, and the number ranged from 6 to 310 per tumour (Table 1).

Treatment of the KPC tumours with BH resulted in the formation of hyperechoic regions seen using ultrasound imaging in all except one tumour. These hyperechoic signals were seen in the core of the tumours where the exposures were targeted, and no such signals were detected outside the tumours (Figure 1). Treatment delivery resulted in no side effects such as decreased quality of life, signs of pain, or the development of damage to the overlying skin, peritoneum, or other adjacent organs in subjects as assessed at the time of dissection. Despite these treatments, all tumours (control, reovirus, BH, combination groups) continued to grow, with the subjects ultimately succumbing to their disease.

3.2. Histological Analysis Following Treatments

Sham-exposed and reovirus-treated tumours had densely packed KPC cells covering their core and periphery 6 and 72 h after treatment (Figure 2A–C). Occasionally, necrotic-looking areas characterised by a reduced number of cell nuclei (Figure 2A1,B2) and haematomas (Figure 2A) could be seen. When tumours were exposed to BH, either alone or in the combination group, extensive tissue disruption was seen 6 h after treatment (Figure 2D and Figure 2E, respectively). These ultrasound-treated tissues exhibited morphological tissue structure features including increased intercellular gaps, and abnormal-looking and elongated nuclei (Figure 2D1,E2). In some cases, relatively large haematomas were observed (Figure 2E1). Seventy-two (72) hours after combination and BH-only exposures, extensive necrotic-like regions were seen (Figure 2F,G). These areas had markedly different nuclear and cytoplasmic staining compared to the sham-exposed and reovirus-treated tissues, as well as compared to the normal-looking KPC peripheral tissue of the same tumours (Figure 2F1,G1). These differences included eosin-stained intercellular gaps possibly consisting of debris and fibres in the treated areas and a low number of cell nuclei (Figure 2F2,G2).

3.3. Characterisation of Immune Subset Populations

The two-dimensional transcriptomic profile of the CD45⁺ immune cell populations (leukocytes) found in the tumours of the four groups (sham controls, BH, reovirus, and combination) 24 h after treatment corresponded to 30 distinct clusters in a UMAP (Figure 3) which overlapped with the protein staining of the AbSeq antibodies as shown in Figure A2. The clusters were further annotated manually using biomarkers commonly associated with each immune cell type (as listed in

Table 2. Manual annotation of immune cell clusters observed in KPC tumours. Thirty (30) cell clusters were identified in the UMAP of the KPC tumours used in this study. These clusters were annotated after assessing conserved gene expression within each cluster. These conserved genes are listed for neutrophils (Table 2A,B), CD4⁺ lymphocytes (Table 2C), and CD8⁺ lymphocytes (Table 2D). Gene overexpression in a cluster was considered “high” if their AvgLog₂Fc > 1 (dark green), “moderate” if 0.5 < AvgLog₂Fc < 1 (light green), “low” if 0.25 < AvgLog₂FC < 0.5 (amber), and “negative” (neg) if AvgLog₂FC < 0.25 (red).

A—Neutrophils

Clusters

Relative expression of annotation genes—AvgLog2Fc > 1 (dark green), 0.5 < AvgLog2Fc < 1 (light green), 0.25 < AvgLog2FC < 0.5 (amber), AvgLog2FC < 0.25 (red).

Transitioning

CD63

ITGAM

FCGR3

SELL

ENTPD1

PDL1

CYBB

CD177

CXCR2

MMP9

IL10R

OSM

TGF-βR

IFNγ/IFNγR

TLR4/CD14

TLR2

CCL2

S100A8 S100A9

TNF

CD24A

Active

CD63

ITGAM

FCGR3

SELL

ENTPD1

PDL1

CYBB

CD177

CXCR2

MMP9

IL10R

OSM

TGF-βR

IFNγ/IFNγR

TLR4/CD14

TLR2

CCL2

S100A8 S100A9

TNF

CD24A

Active subset

CD63

ITGAM

FCGR3

SELL

ENTPD1

PDL1

CYBB

CD177

CXCR2

MMP9

IL10R

OSM

TGF-βR

IFNγ/IFNγR

TLR4/CD14

TLR2

CCL2

S100A8 S100A9

TNF

CD24A

Intermediate

CD63

ITGAM

FCGR3

SELL

ENTPD1

PDL1

CYBB

CD177

CXCR2

MMP9

IL10R

OSM

TGF-βR

IFNγ/IFNγR

TLR4/CD14

TLR2

CCL2

S100A8 S100A9

TNF

CD24A

Intermediate subset

CD63

ITGAM

FCGR3

SELL

ENTPD1

PDL1

CYBB

CD177

CXCR2

MMP9

IL10R

OSM

TGF-βR

IFNγ/IFNγR

TLR4/CD14

TLR2

CCL2

S100A8 S100A9

TNF

CD24A

Inactive

CD63

ITGAM

FCGR3

SELL

ENTPD1

PDL1

CYBB

CD177

CXCR2

MMP9

IL10R

OSM

TGFbR

IFNγ/IFNγR

TLR4/CD14

TLR2

CCL2

S100A8 S100A9

TNF

CD24A

Inactive subset

CD63

ITGAM

FCGR3

SELL

ENTPD1

PDL1

CYBB

CD177

CXCR2

MMP9

IL10R

OSM

TGF-βR

IFNγ/IFNγR

TLR4/CD4

TLR2

CCL2

S100A8 S100A9

TNF

CD24A

B—Neutrophils

Clusters

Additional annotation genes. Relative expression—AvgLog2Fc > 1 (dark green), 0.5 < AvgLog2Fc < 1 (light green), 0.25 < AvgLog2FC < 0.5 (amber), AvgLog2FC < 0.25 (red).

Active

MMP8

IFIT2

IFITM6

CXCL10

CXCL3

CCL4

CCL6

IFIT3

IFIT3B

IFITM3

IFIT1

ISG20

IFIT1BL2

TNFAIP6

DACH1

MRGPRA2B

GSTM1

IL23A

SIGLECF

LTF

ADAM8

ADAMDEC1

Active subset

MMP8

IFIT2

IFITM6

CXCL10

CXCL3

CCL4

CCL6

IFIT3

IFIT3B

IFITM3

IFIT1

ISG20

IFIT1BL2

TNFAIP6

DACH1

MRGPRA2B

GSTM1

IL23A

SIGLECF

LTF

ADAM8

ADAMDEC1

Intermediate

MMP8

IFIT2

IFITM6

CXCL10

CXCL3

CCL4

CCL6

IFIT3

IFIT3B

IFITM3

IFIT1

ISG20

IFIT1BL2

TNFAIP6

DACH1

MRGPRA2B

GSTM1

IL23A

SIGLECF

LTF

ADAM8

ADAMDEC1

Intermediate subset

MMP8

IFIT2

IFITM6

CXCL10

CXCL3

CCL4

CCL6

IFIT3

IFIT3B

IFITM3

IFIT1

ISG20

IFIT1BL2

TNFAIP6

DACH1

MRGPRA2B

GSTM1

IL23A

SIGLECF

LTF

ADAM8

ADAMDEC1

Transitioning

MMP8

IFIT2

IFITM6

CXCL10

CXCL3

CCL4

CCL6

IFIT3

IFIT3B

IFITM3

IFIT1

ISG20

IFIT1BL2

TNFAIP6

DACH1

MRGPRA2B

GSTM1

IL23A

SIGLECF

LTF

ADAM8

ADAMDEC1

Inactive

MMP8

IFIT2

IFITM6

CXCL10

CXCL3

CCL4

CCL6

IFIT3

IFIT3B

IFITM3

IFIT1

ISG20

IFIT1BL2

TNFAIP6

DACH1

MRGPRA2B

GSTM1

IL23A

SIGLECF

LTF

ADAM8

ADAMDEC1

Inactive subset

MMP8

IFIT2

IFITM6

CXCL10

CXCL3

CCL4

CCL6

IFIT3

IFIT3B

IFITM3

IFIT1

ISG20

IFIT1BL2

TNFAIP6

DACH1

MRGPRA2B

GSTM1

IL23A

SIGLECF

LTF

ADAM8

ADAMDEC1

C—CD4+ lymphocytes—Relative expression of annotation genes—AvgLog2Fc > 1 (dark green), 0.5 < AvgLog2Fc < 1 (light green), 0.25 < AvgLog2FC < 0.5 (amber), AvgLog2FC < 0.25 (red).

Clusters

Marker genes

Activation genes

Immune checkpoint genes

Proliferation genes

Various genes

Naïve/ inactive Th

FOXP3

CD3e

CD4

FR4

CD25

CD69

TNFRSF9

TNFRSF4

CCR5

IFNγ

CTLA4

PDCD1

LAG3

TIM3

TIGIT

ICOS

MKI67

TOP2A

PCNA

CD40LG

TBET

IL7R

Active Th

FOXP3

CD3e

CD4

FR4

CD25

CD69

TNFRSF9

TNFRSF4

CCR5

IFNγ

CTLA4

PDCD1

LAG3

TIM3

TIGIT

ICOS

MKI67

TOP2A

PCNA

CD40LG

TBET

IL7R

Proliferating active Th

FOXP3

CD3e

CD4

FR4

CD25

CD69

TNFRSF9

TNFRSF4

CCR5

IFNγ

CTLA4

PDCD1

LAG3

TIM3

TIGIT

ICOS

MKI67

TOP2A

PCNA

CD40LG

TBET

IL7R

Treg

FOXP3

CD3e

CD4

FR4

CD25

CD69

TNFRSF9

TNFRSF4

CCR5

IL10

CTLA4

PDCD1

LAG3

TIM3

TIGIT

ICOS

MKI67

TOP2A

PCNA

CD40LG

KLRG1/CCR2

ITGAE

IL7R

Tfh

FOXP3

CD3g

CD4

FR4

CD25

CD69

TNFRSF9

TNFRSF4

CCR5

IL10

CTLA4

PDCD1

LAG3

TIM3

TIGIT

ICOS

MKI67

TOP2A

CCNA2

CD40LG

IL10/FCGR3

TNF

IL7R

D—CD8+ lymphocytes—Relative expression of annotation genes—AvgLog2Fc > 1 (dark green), 0.5 < AvgLog2Fc < 1 (light green), 0.25 < AvgLog2FC < 0.5 (amber), AvgLog2FC < 0.25 (red).

Clusters

Lineage genes

Activation genes

Immune checkpoint genes

Proliferation/cell cycle genes

Transcription factor genes

Various genes

Senescent Tcyt

KLRG1

CD3e

CD8α

GZMB

FASL

CD69

TNFRSF9

PFN

LAG3

PDCD1

CTLA4

TIGIT

TIM3

MKI67

TOP2A

PCNA

CCNA2

TBET

TOX

EOMES

IFNγ

SELL

IL7R

CD44

Exhausted proliferating Tcyt

KLRG

CD3e

CD8α

GZMB

FASL

CD69

TNFRSF9

PFN

LAG3

PDCD1

CTLA4

TIGIT

TIM3

MKI67

TOP2A

PCNA

CCNA2

TBET

TOX

EOMES

IFN-γ

SELL

IL7R

CD44

Terminally Exhausted Tcyt

KLRG

CD3e

CD8α

GZMB

FASL

CD69

TNFRSF9

PFN

LAG3

PDCD1

CTLA4

TIGIT

TIM3

MKI67

TOP2A

PCNA

CCNA2

TBET

TOX

EOMES

IFNγ

SELL

IL7R

CD44

Exhausted Tcyt

KLRG

CD3e

CD8α

GZMB

FASL

CD69

TNFRSF9

PFN

LAG3

PDCD1

CTLA4

TIGIT

TIM3

MKI67

TOP2A

PCNA

CCNA2

TBET

TOX

EOMES

IFNγ

SELL

IL7R

CD44

Tcyt subset 1

KLRG

CD3e

CD8α

GZMB

FASL

CD69

TNFTSF9

PFN

LAG3

PDCD1

CTLA4

TIGIT

TIM3

MKI67

TOP2A

PCNA

CCNA2

TBET

TOX

EOMES

IFNγ

SELL

IL7R

CD44

Tcyt subset 2

KLRG

CD3e

CD8α

GZMB

FASL

CD69

TNFTSF9

PFN

LAG3

PDCD1

CTLA4

TIGIT

TIM3

MKI67

TOP2A

PCNA

CCNA2

TBET

TOX

EOMES

IFNγ

SELL

IL7R

CD44

Proliferating Tcyt

KLRG

CD3e

CD8α

GZMB

FASL

CD69

TNFTSF9

PFN

LAG3

PDCD1

CTLA4

TIGIT

TIM3

MKI67

TOP2A

PCNA

CCNA2

TBET

TOX

EOMES

IFNγ

SELL

IL7R

CD44

Non-proliferating Tcyt

KLRG

CD3e

CD8α

GZMB

FASL

CD69

TNFTSF9

PFN

LAG3

PDCD1

CTLA4

TIGIT

TIM3

MKI67

TOP2A

PCNA

CCNA2

TBET

TOX

EOMES

IFNγ

SELL

IL7R

CD44

TRM CD8T

ITGAE

CD3e

CD8α

GZMB

FASL

CD69

TNFRSF9

PFN

LAG3

PDCD1

CTLA4

TIGIT

TIM3

MKI67

TOP2A

PCNA

CCNA2

TBET

TOX

EOMES

ICOS

IL2RB

IL7R

ITGA1

A–D,

Table A2. Differences in conserved gene expression of exhausted T cell subsets. Gene overexpression in a cluster was considered “high” if their AvgLog₂Fc > 1 (dark green), “moderate” if 0.5 < AvgLog₂Fc < 1 (light green), “low” if 0.25 < AvgLog₂FC < 0.5 (amber), and “negative” (neg) if AvgLog₂FC < 0.25 (red).

CD8+ lymphocytes
Clusters	Additional annotation genes—Relative expression—AvgLog2Fc > 1 (dark green), 0.5 < AvgLog2Fc < 1 (light green), 0.25 < AvgLog2FC < 0.5 (amber), AvgLog2FC < 0.25 (red).
Terminally Exhausted Tcyt	MZB1	DERL3	CRELD2	CDK8	PIK3CG	Ribosomal proteins	SLPI	SPP1	AIM1	TCRG.C4
Exhausted Tcyt	MZB1	DERL3	CRELD2	CDK8	PIK3CG		SLP1	SPP1	AIM1	TCRG.C4

and

Table A3. Manual annotation of NK, DC, B cells, macrophages/monocytes, damaged cells and SIRPα⁺ cells clusters observed in KPC tumours. Clusters were annotated after assessing conserved gene expression within each cluster. Gene overexpression in a cluster was considered “high” if their AvgLog₂Fc > 1 (dark green), “moderate” if 0.5 < AvgLog₂Fc < 1 (light green), “low” if 0.25 < AvgLog₂FC < 0.5 (amber), and “negative” (neg) if AvgLog₂FC < 0.25 (red).

Natural killer, dendritic, B cells, macrophages/monocytes, damaged, MDSC and SIRPα+ cells—Relative expression of annotation genes- AvgLog2Fc > 1 (dark green), 0.5 < AvgLog2Fc < 1 (light green), 0.25 < AvgLog2FC < 0.5 (amber), AvgLog2FC < 0.25 (red).

Clusters

Annotation genes

NK

KLRB1

CD3e

CD8a

GZMB

FASL

KLRK1

KLRC1

IL7R

TBET

EOMES

DC

ITGAX

CLEC9A

XCR1

CD40

CD86

CD83

TIM3

CCR7

ITGAE

B cells

CD24a

CD38

CD19

CCR7

CD69

MS4A1

IGHD

CD86

CR2

CD69

CYBB

B transitioning

CD24a

CD38

CD19

CCR7

CD69

MS4A1

IGHD

CD86

CR2

CD69

CYBB

Macrophages

ITGAX

ITGB2

CD86

ADGRE1

CD68

ARG1

FCGR1

FCGR3

ITGAM

CCR2

CSF1R

CYBB

CCR5

MRC1

CCL2

IFITM1

S100A8

Inflammatory Monocytes

CD14

iNOS

SELL

TNFRSF1α

ITGB2

CD86

ADGRE1

CD68

ARG1

FCGR1

FCGR3

ITGAM

CCR2

CSF1R

CYBB

CCR5

MRC1

CCL2

IFITM1

S100A8

MDSC-enriched

CD33

CD14

iNOS

SELL

PDL1

ADGRE1

CD68

ARG1

FCGR1

FCGR3

ITGAM

CCR2

CSF1R

CYBB

RETNLG

MRC1

CCL2

IFITM1

S100A8

Damaged cells

MALAT1

Mitochondrial genes

SIRPα+ cells

ADGRE1

CD3e

CD68

CD44

ITGAX

CD86

CD14

CD16

SELL

CD8α

ITGAL

FAS

CTLA4

). Neutrophils (seven clusters) (Table 2A,B), CD4⁺ lymphocytes (Th) (five clusters) (Table 2C), nine CD8⁺ lymphocytes clusters (Table 2D and Table A2) and two B cell clusters and single clusters of macrophages, inflammatory monocytes, SIRPα⁺ macrophages, natural killer (NK) cells, dendritic cells (DC), myeloid-derived suppressor cells (MDSC), and “leaky” damaged cells (Table A3) were identified.

This detailed analysis showed that some of these cell subsets were associated with an anti-immune pro-tumour gene expression phenotype. For example, all neutrophil subtypes (active, intermediate, inactive, transitioning neutrophils including subsets) overexpressed anti-immune genes associated with an N2-like phenotype, e.g., CXCR2, IL10R, CD177, OSM, TGF-βR (all but the active neutrophils subset cluster) and PDL1 (Table 2A). Cytotoxic T cell (Tcyt) clusters including the senescent Tcyt, terminally exhausted Tcyt, exhausted Tcyt, exhausted proliferating Tcyt, as well as the tissue-resident memory (TRM) CD8⁺T overexpressed one or more of the immune checkpoint genes LAG3, PDCD1, CTLA4, TIGIT, and TIM3 (Table 2D). A small but distinct set of macrophages showed significant gene downregulation (~76% of all its genes) and stained positive for the SIRPα protein (Table A3 and Figure A2L) suggesting the prior activation of the CD47-SIRPα signalling axis in this macrophage subset.

3.4. Relative Immune Cell Abundance Following Treatment

The relative abundance of the 30 immune cell subtypes (compared to the total leukocyte abundance) in the tumours 24 h after sham, BH, reovirus, and combination treatments is shown in Figure 4, Figure A3 and Figure A4A and

Table 3. Relative abundance of immune cells in PDAC tumours 24 h after treatment. The percentages of each cell type relative to the CD45⁺ cells in the tumours of sham-exposed “control” subjects (n = 6), reovirus-treated group (n = 6), BH-treated group (n = 5), and the combination-treated group (n = 4) are shown. Percentages are shown as averages ± SEM. Statistical significance compared to controls is assumed at p < 0.05, and these values are highlighted in red.

Cell Abundance as a % of CD45+ Cells	Control (n = 6)	Reovirus (n = 6)	BH (n = 5)	Combination (n = 4)
Macrophages	12.20 ± 1.4	7.75 ± 1.3	12.21 ± 2.1	6.94 ± 1.4
Senescent Tcyt	8.30 ± 1.9	7.67 ± 1.4	4.78 ± 1.1	4.96 ± 0.9
Active neutrophils	4.86 ± 1.6	6.76 ± 2.3	10.04 ± 3.3	13.81 ± 6.5
Exhausted proliferating Tcyt	7.18 ± 2.2	7.166 ± 3.5	5.99 ± 3.03	8.49 ± 3.9
Terminal exhausted Tcyt	5.58 ± 1.2	5.41 ± 1.2	5.12 ± 0.8	3.26 ± 0.5
Active Th	6.11 ± 1.4	5.82 ± 1.2	5.08 ± 0.4	5.29 ± 1.2
Intermediate neutrophils	4.62 ± 1.0	5.12 ± 1.6	6.81 ± 1.6	8.90 ± 2.9
Inflammatory monocytes	4.59 ± 1.4	5.28 ± 0.7	5.63 ± 0.6	7.60 ± 1.8
TRM CD8T	6.04 ± 0.8	4.73 ± 1.2	4.11 ± 0.7	3.25 ± 1.1
Proliferating active Th	5.42 ± 0.6	4.82 ± 0.9	4.14 ± 0.7	3.08 ± 0.2
Tcyt subset1	4.46 ± 1.7	4.60 ± 1.8	4.71 ± 1.4	2.20 ± 1.1
B cells	3.60 ± 1.1	4.49 ± 2.0	7.25 ± 4.3	1.34 ± 0.3
Inactive neutrophils	3.96 ± 3.0	5.22 ± 4.5	2.07 ± 1.5	7.15 ± 4.1
DC	3.16 ± 0.2	3.76 ± 0.8	3.15 ± 0.7	2.67 ± 0.6
Exhausted Tcyt	2.64 ± 0.5	2.79 ± 0.6	3.11 ± 0.5	1.67 ± 0.3
Damaged cells	1.98 ± 0.5	2.39 ± 0.6	2.79 ± 0.5	2.42 ± 0.4
NK	2.00 ± 0.5	1.91 ± 0.4	2.24 ± 0.6	1.20 ± 0.2
Treg	2.29 ± 0.3	2.19 ± 0.5	1.23 ± 0.2	0.80 ± 0.2
Transitioning neutrophils	1.35 ± 0.2	1.98 ± 0.9	2.81 ± 0.6	3.03 ± 1.1
Tcyt subset2	1.72 ± 0.6	1.59 ± 0.4	1.91 ± 0.4	0.90 ± 0.3
Naïve/inactive Th	1.52 ± 0.5	1.66 ± 0.6	1.38 ± 0.2	0.87 ± 0.4
Inactive neutrophils subset	1.54 ± 1.3	1.73 ± 1.6	0.69 ± 0.6	2.07 ± 1.4
Proliferating Tcyt	1.13 ± 0.1	1.39 ± 0.2	0.72 ± 0.1	0.87 ± 0.3
MDSC-enriched	0.81 ± 0.2	0.83 ± 0.4	1.39 ± 0.5	1.46 ± 0.6
Active neutrophils subset	0.61 ± 0.5	1.02 ± 0.9	0.43 ± 0.4	3.36 ± 1.9
Intermediate neutrophils subset	0.87 ± 0.7	0.83 ± 0.8	0.23 ± 0.2	1.14 ± 0.6
B transitioning	0.80 ± 0.6	0.41 ± 0.3	0.33 ± 0.4	0.45 ± 0.4
Non-proliferating Tcyt	0.29 ± 0.2	0.27 ± 0.2	0.39 ± 0.4	0.38 ± 0.3
SIRPa+ cells	0.19 ± 0.1	0.18 ± 0.1	0.12 ± 0.1	0.21 ± 0.1
Tfh	0.04 ± 0.1	0.08 ± 0.1	0.03 ± 0.1	0.09 ± 0.1

. The statistical significance of the results for each cell type in the treatment groups compared to the sham-exposed subjects is shown in

Table A4. Statistical significance (p-values) of the differences in relative cell abundance between sham-exposed controls and each treatment group (Table 3) are presented. Statistical significance is assumed at p < 0.05, and these values are highlighted in red.

p-Values (Cell Abundance)	Control vs. Reovirus	Control vs. BH	Control vs. Combination
Macrophages	0.019	0.997	0.019
Senescent Tcyt	0.759	0.110	0.158
Active neutrophils	0.423	0.123	0.135
Exhausted proliferating Tcyt	0.996	0.718	0.741
Terminal exhausted Tcyt	0.904	0.730	0.112
Active Th	0.844	0.437	0.635
Intermediate neutrophils	0.759	0.193	0.125
Inflammatory monocytes	0.604	0.470	0.174
TRM CD8T	0.286	0.062	0.046
Proliferating active Th	0.515	0.153	0.007
Tcyt subset1	0.945	0.896	0.273
B cells	0.647	0.344	0.095
Inactive neutrophils	0.784	0.546	0.497
DC	0.366	0.991	0.365
Exhausted Tcyt	0.833	0.520	0.199
Damaged cells	0.546	0.188	0.470
NK	0.855	0.728	0.147
Treg	0.823	0.016	0.004
Transitioning neutrophils	0.426	0.027	0.089
Tcyt subset2	0.838	0.763	0.258
Naïve/inactive Th	0.819	0.769	0.285
Inactive neutrophils subset	0.907	0.513	0.758
Proliferating Tcyt	0.225	0.016	0.310
MDSC-enriched	0.951	0.220	0.268
Active neutrophils subset	0.625	0.750	0.127
Intermediate neutrophils subset	0.967	0.356	0.759
B transitioning	0.515	0.483	0.629
Non-proliferating Tcyt	0.939	0.813	0.771
SIRPa+ cells	0.947	0.632	0.928
Tfh	0.221	0.720	0.371

. Statistically significant increases were seen in the abundance of the transitioning neutrophils (2.1-fold) for the BH group compared to that in the control group (Figure 4A), as well as for the aggregated active neutrophils (3.1-fold) (Figure 4A), total neutrophils (active, intermediate, inactive, and transitioning clusters and subsets) (2.2-fold) (Figure 4A), and the total neutrophil to total lymphocyte (CD8 T cells, CD4 T cells, NK, and B cells) ratio (NLR) (3.2-fold) (Figure A3) in the tumours of the combination group compared to the sham-exposed subjects. When the reovirus group and the BH group were compared to the sham-exposed group, no statistically significant differences were seen in the neutrophil abundance or the NLR.

Statistically significant decreases in the abundance of macrophages (1.8-fold) (Figure 4B), proliferating active Th (1.8-fold) (Figure 4C), regulatory T cells (Treg) (2.9-fold) (Figure 4C), TRM CD8T (1.9-fold) (Figure 4D), and total B cells (B cells + B transitioning) (2.4-fold) (Figure A4A) were seen in the combination-treated subjects compared to the sham-exposed ones. Statistically significant decreases were also seen for the Treg (1.9-fold) (Figure 4C) and proliferating Tcyt (1.6-fold) (Figure 4D) of the BH-exposed group and for the macrophages (1.6-fold) of the reovirus treated group (Figure 4B) compared with the sham-exposed group.

Ultrasound exposures (BH and combination treatments) increased the balance of immune cells to Treg cells compared to the sham-exposed group (Figure 5 and Figure A4B and

Table 4. Ratio of immune cells to Treg 24 h after treatment. The ratios of the abundance of each immune cell type to that of Treg in the KPC tumours of each treatment group are shown as averages ± SEM. Statistical significance compared to controls is assumed at p < 0.05, and these values are highlighted in red.

Immune Cell Ratio to Treg	Control (n = 6)	Reovirus (n = 6)	BH (n = 5)	Combination (n = 4)
Macrophages	5.31 ± 0.61	3.52 ± 0.62	9.93 ± 1.73	8.67 ± 1.73
Senescent Tcyt	3.61 ± 0.85	3.48 ± 0.66	3.89 ± 0.89	6.20 ± 1.20
Active neutrophils	2.11 ± 0.71	3.07 ± 1.02	8.16 ± 2.67	17.27 ± 8.11
Exhausted proliferating Tcyt	3.12 ± 0.97	3.25 ± 1.57	4.87 ± 2.46	10.61 ± 4.85
Terminal exhausted Tcyt	2.42 ± 0.53	2.46 ± 0.52	4.16 ± 0.66	4.07 ± 0.61
Active Th	2.65 ± 0.59	2.64 ± 0.52	4.13 ± 0.30	6.61 ± 1.46
Intermediate neutrophils	2.01 ± 0.44	2.32 ± 0.74	5.54 ± 1.27	11.13 ± 3.60
Inflammatory monocytes	1.99 ± 0.62	2.40 ± 0.31	4.58 ± 0.49	9.50 ± 2.20
TRM CD8T	2.62 ± 0.34	2.15 ± 0.54	3.34 ± 0.57	4.06 ± 1.38
Proliferating active Th	2.35 ± 0.27	2.19 ± 0.40	3.36 ± 0.60	3.85 ± 0.23
Tcyt subset1	1.94 ± 0.72	2.09 ± 0.80	3.83 ± 1.09	2.75 ± 1.42
B cells	1.56 ± 0.49	2.04 ± 0.91	5.89 ± 3.5	1.68 ± 0.42
Inactive neutrophils	1.72 ± 1.33	2.37 ± 2.07	1.68 ± 1.23	8.94 ± 5.10
DC	1.37 ± 0.09	1.71 ± 0.34	2.56 ± 0.56	3.34 ± 0.72
Exhausted Tcyt	1.14 ± 0.28	1.27 ± 0.27	2.53 ± 0.40	2.09 ± 0.32
Damaged cells	0.86 ± 0.20	1.08 ± 0.29	2.27 ± 0.37	3.03 ± 0.52
NK	0.87 ± 0.20	0.87 ± 0.17	1.82 ± 0.49	1.50 ± 0.20
Transitioning neutrophils	0.58 ± 0.09	0.90 ± 0.41	2.28 ± 0.52	3.78 ± 1.33
Tcyt subset2	0.74 ± 0.27	0.72 ± 0.20	1.56 ± 0.30	1.12 ± 0.34
Naïve/inactive Th	0.66 ± 0.21	0.75 ± 0.26	1.12 ± 0.16	1.09 ± 0.44
Inactive neutrophils subset	0.66 ± 0.54	0.79 ± 0.72	0.56 ± 0.50	2.59 ± 1.76
Proliferating Tcyt	0.49 ± 0.05	0.63 ± 0.09	0.58 ± 0.07	1.09 ± 0.32
MDSC-enriched	0.35 ± 0.08	0.37 ± 0.16	1.13 ± 0.41	1.82 ± 0.81
Active neutrophils subset	0.26 ± 0.21	0.46 ± 0.39	0.35 ± 0.30	4.20 ± 2.43
Intermediate neutrophils subset	0.37 ± 0.30	0.37 ± 0.35	0.19 ± 0.19	1.42 ± 0.76
B transitioning	0.34 ± 0.27	0.18 ± 0.14	0.27 ± 0.31	0.56 ± 0.43
Non-proliferating Tcyt	0.12 ± 0.09	0.12 ± 0.09	0.31 ± 0.35	0.47 ± 0.31
SIRPa+ cells	0.08 ± 0.06	0.08 ± 0.05	0.10 ± 0.06	0.26 ± 0.12
Tfh	0.02 ± 0.01	0.04 ± 0.01	0.03 ± 0.01	0.11 ± 0.07

and

Table A5. Statistical significance of the differences in the immune cell/Treg ratio between sham-exposed controls and each treatment group (Table 4) are presented. Statistical significance is assumed at p < 0.05, and these values are highlighted in red.

p-Values (Cell Ratios over Treg)	Control vs. Reovirus	Control vs. BH	Control vs. Combination
Macrophages	0.031	0.0132	0.056
Senescent Tcyt	0.892	0.798	0.079
Active neutrophils	0.368	0.025	0.046
Exhausted proliferating Tcyt	0.930	0.441	0.095
Terminal exhausted Tcyt	0.957	0.038	0.053
Active Th	0.984	0.034	0.016
Intermediate neutrophils	0.664	0.010	0.013
Inflammatory monocytes	0.491	0.004	0.003
TRM CD8T	0.383	0.231	0.244
Proliferating active Th	0.688	0.097	0.002
Tcyt subset1	0.866	0.119	0.560
B cells	0.587	0.168	0.850
Inactive neutrophils	0.753	0.979	0.128
DC	0.27	0.030	0.009
Exhausted Tcyt	0.709	0.008	0.039
Damaged cells	0.461	0.003	0.001
NK	0.990	0.058	0.041
Transitioning neutrophils	0.385	0.003	0.016
Tcyt subset2	0.928	0.043	0.361
Naïve/inactive Th	0.735	0.074	0.317
Inactive neutrophils subset	0.874	0.876	0.234
Proliferating Tcyt	0.152	0.252	0.049
MDSC-enriched	0.866	0.047	0.051
Active neutrophils subset	0.600	0.780	0.076
Intermediate neutrophils subset	0.997	0.566	0.160
B transitioning	0.539	0.833	0.631
Non-proliferating Tcyt	0.978	0.536	0.219
SIRPa+ cells	0.991	0.841	0.161
Tfh	0.188	0.554	0.120

). For the BH and combination groups compared to the sham-exposed controls, statistically significant increases were seen for the intermediate neutrophils (2.8- and 5.5-fold, respectively) (Figure 5A), transitioning neutrophils (3.9- and 6.4-fold) (Figure 5A), inflammatory monocytes (2.3- and 4.8-fold) (Figure 5B), DC (1.9- and 2.4-fold) (Figure 5B), active Th (1.6- and 2.5-fold) (Figure 5C), and exhausted Tcyt (2.2- and 1.8-fold) (Figure 5D). In addition, for the combination vs. sham-exposed groups only increases were seen for the ratios of the NK cells (1.7-fold) (Figure 5B), proliferating active Th (1.6-fold) (Figure 5C), and proliferating Tcyt (2.2-fold) (Figure 5D) to Treg. For the BH-exposed tumours, increases were seen for the ratios of the macrophages (1.9-fold) (Figure 5B), MDSC-enriched cells (3.2-fold) (Figure 5B), terminal exhausted Tcyt (1.7-fold) (Figure 5D), and Tcyt subset2 (2.1-fold) (Figure 5D) over Treg. In the reovirus group, a decrease in the macrophages/Treg (1.5-fold) (Figure 5B) was seen compared to the sham-exposed subjects.

Statistically significant differences in the ratio of immune cells to Treg were also seen between treatment groups. Compared to the reovirus-treated subjects, the combined treatments increased the ratios of the intermediate neutrophils to Treg (4.8-fold), and transitioning neutrophils (4.2-fold) (Figure 5A), inflammatory monocytes (4.0-fold), macrophages (2.5-fold), NK cells (1.7-fold) (Figure 5B), active Th (2.5-fold) (Figure 5C), and senescent Tcyt (1.8-fold) (Figure 5D). Also, the combined treatments increased the ratio of the inflammatory monocytes to Treg (2.1-fold) when compared to the BH-only exposures (Figure 5B). The BH exposures, compared to the reovirus-only treatments, resulted in increased ratios of intermediate neutrophils to Treg (2.4-fold), and transitioning neutrophils (2.5-fold) (Figure 5A), macrophages (2.8-fold), inflammatory monocytes (1.9-fold) (Figure 5B), active Th (1.6-fold) (Figure 5C), terminal exhausted Tcyt (1.7-fold), exhausted Tcyt (2.0-fold), and Tcyt subset2 (2.2-fold) (Figure 5D).

3.5. Effects of Treatments on Differential Gene Expression in All Immune Cell Clusters

Differential gene expression (DEG) in all clusters treated with reovirus, BH, BH + reovirus, and in sham-exposed controls 24 h after treatment was investigated to characterise in detail the transcriptomic profile of the tumoural immune cell infiltrates.

Compared to the sham-exposed controls, reovirus treatments differentially modulated 655 genes in all clusters (

Table 5. The number of differentially expressed immune cell genes in this study. The number of genes upregulated or downregulated are shown for every immune cell subtype for comparisons of control vs. reovirus, control vs. BH, control vs. combination, combination vs. reovirus, combination vs. BH, and reovirus vs. BH. Genes were assumed to be significantly regulated if their AvgLog₂FC expression > 1 and their p-adjusted value < 0.05.

Cell Clusters	Number of Genes
	Control vs. Reovirus		Control vs. BH		Control vs. Combination		Combination vs. Reovirus		Comb. vs. BH		Reovirus vs. BH
	Virus Up	Virus Down	BH Up	BH Down	Combination Up	Combination Down	Virus Up	Virus Down	BH Up	BH Down	BH Up	BH Down
Senescent Tcyt	2	0	0	0	3	1	1	0	0	3	1	0
Exhausted prolifer. Tcyt	18	0	30	6	15	1	1	1	16	63	153	23
Terminal exhausted Tcyt	7	0	0	0	4	1	1	0	0	0	0	0
Exhausted Tcyt	10	1	0	1	1	5	3	0	7	15	3	4
Tcyt subset1	0	0	0	0	5	2	2	2	1	2	0	0
Tcyt subset2	1	0	5	2	2	1	1	1	1	1	4	3
Proliferating Tcyt	1	0	0	0	0	0	0	0	0	0	0	0
Non-proliferating Tcyt	0	0	0	0	1	0	0	1	0	0	0	0
TRM CD8T	1	0	0	0	2	0	0	1	0	0	0	1
NK	6	0	0	0	3	6	0	2	1	0	3	1
DC	0	0	0	38	0	0	0	0	1	1	0	0
B cells	11	0	0	1	11	2	1	30	23	0	0	0
B transitioning	0	0	0	0	11	2	0	0	16	0	0	0
Naïve/inactive Th	2	1	0	0	3	0	0	0	1	1	0	0
Active Th	34	4	0	0	18	5	0	1	14	3	0	0
Proliferating active Th	32	3	0	0	34	17	0	5	12	11	1	0
Treg	2	0	0	0	1	1	1	0	0	1	0	0
Tfh	0	0	0	0	0	0	0	0	0	0	0	0
Macrophages	152	73	15	105	97	49	0	0	36	52	72	205
Inflammatory monocytes	130	94	0	0	165	34	0	1	83	43	109	205
MDSC enriched	9	4	1	0	89	17	4	31	8	4	0	0
SIRPα+ cells	3	0	2	0	3	0	3	0	18	3	0	0
Active neutrophils	37	1	151	17	408	217	0	1	12	1	4	11
Active neutrophils subset	0	0	1	0	1	0	0	0	0	0	0	0
Intermediate neutrophils	1	0	176	21	451	190	55	155	11	4	40	13
Intermed. Neutroph. subset	0	0	0	0	0	1	0	0	0	0	0	0
Transitioning neutrophils	15	0	9	9	131	51	2	13	10	1	0	4
Inactive neutrophils	0	0	17	1	203	325	111	58	26	49	2	1
Inactive neutrophils subset	0	0	0	0	31	44	8	15	4	0	0	0
Total number of genes differentially regulated	474	181	407	201	1693	972	194	318	301	258	392	471
	655		608		2665		512		559		863

), causing significant DEG in macrophages and inflammatory monocytes (>200 genes differentially modulated in each cell type), active Th and proliferating active Th (≥35 genes differentially modulated in each cell type), exhausted proliferating Tcyt (18 genes upregulated), and active neutrophils (37 genes upregulated). BH treatments alone modulated the expression of 608 genes (Table 5), particularly in active and intermediate neutrophils (>150 genes upregulated in each cell type) and in the exhausted proliferating Tcyt (30 genes upregulated). Unlike the reovirus treatments, BH caused no significant changes in the gene expression of inflammatory monocytes and downregulated 105 macrophage and 38 DC genes compared to control subjects (Table 5).

Compared to the sham-exposed controls, the combined treatments differentially modulated more genes (>2-fold) than the reovirus-alone and BH-alone treatments combined. In total, the transcription of ~2600 genes was affected among all cell clusters (Table 5) with ~1700 genes significantly increasing their expression and ~950 genes decreasing their expression after the combined treatments. The cell types with the highest DEG were the active, intermediate, and inactive neutrophils (>500 genes differentially modulated in each cell type), the transitioning neutrophils and cells of monocytic origin such as the macrophages and inflammatory monocytes (>140 genes differentially modulated), followed by the Th cells and various Tcyt clusters (Table 5). Also, after the combined treatments, 18 genes were commonly downregulated (Figure 6A), and 78 genes were commonly upregulated (Figure 6B) among the active, intermediate, transitioning, and inactive neutrophils compared to the controls. In addition, the XAF1 and PYDC4 genes were commonly upregulated among the active neutrophils, macrophages, proliferating Th, and exhausted proliferating Tcyt after the combination treatments (Figure 6C).

Heat maps of the top DEG for each subject showed reasonable gene expression homogeneity among subjects within each treatment and control group. For example, for active neutrophils in the sham vs. combination groups, the top DEG in all sham-exposed subjects were grouped, and similarly, the top DEG in all combination-treated subjects were also grouped (Figure A5A). In the control vs. BH active neutrophils heat map, one outlier of the five BH-treated subjects (subject 2 (Table 1)) shared DEG similarities with the controls (Figure A5B), and the same was true for one outlier from the six reovirus-treated subjects in the control vs. reovirus active neutrophils heat maps (Figure A5C).

3.6. Functional Enrichment Analysis in KPC Tumours

3.6.1. Reovirus Treatments vs. Sham-Exposed Controls

In macrophages, inflammatory monocytes, active Th, proliferating active Th, exhausted proliferating Tcyt, and active neutrophils, 24 h after treatment of the KPC tumours with reovirus, overexpressed genes (an example shown for macrophages in Figure A6A) were associated with the innate immune response including type I interferon (IFN) signalling and production, and defence responses to viruses (Figure A7A–F). In addition, overexpressed genes were associated with the complement activation and apoptotic cell clearance in macrophages (Figure A7A), with the regulation of the MDA-5 signalling in active proliferating Th (Figure A7C), and with cytoplasmic vesicles, the receptor for advanced glycation end-products (RAGE) binding and responses to IFN-β in active neutrophils (Figure A7F).

3.6.2. BH Exposures vs. Sham-Exposed Controls

BH treatments caused the overexpression of genes (an example shown for active neutrophils in Figure A6B) associated with the innate immune response, for example, TLR3 signalling, NF-kB signalling, and IFN-β production, in active (Figure 7A1) and intermediate (Figure 7B) neutrophils 24 h after treatment. But here at least some downregulated genes (Figure A6B) were associated with the transporter associated with antigen processing (TAP)-dependent class I major histocompatibility (MHC) complex, the ABC-type peptide antigen transported activity, and T cell-mediated cytotoxicity (Figure 7A2). In macrophages, downregulated genes (Figure A6C) were associated with type II IFN production, NK cell and lymphocyte-mediated immunity, the α/β T cell receptor complex, T cell receptor signalling, and T cell differentiation, migration, and cell adhesion (Figure 7C). In DC, BH treatments downregulated genes (Figure A6D) that were associated with the α/β T cell activation, IL-2 production, T cell costimulation differentiation activation and signalling, and a small number of genes associated with the TAP and class I MHC complexes, and the NK mediated anti-cancer immune response (Figure 7D).

3.6.3. Combination Treatments vs. Sham-Exposed Controls

Twenty-four (24) hours after treatment of the KPC tumours, in active neutrophils, the combined treatments increased the transcription of genes (Figure A6E) associated with the NF-kβ transcription factor activity, TLR3 signalling, IFN-β production, myeloid leukocytes activation, and chemotaxis compared to the sham-exposed controls (Figure 8A1), whereas downregulated active neutrophil genes (Figure A6E) were associated with the TAP1 complex binding, antigen processing and presentation, and the regulation of T cell-mediated immunity (Figure 8A2). Some of these pathways were also modulated in the other major neutrophil clusters including the intermediate, transitioning, and inactive neutrophils (Figure A8). In macrophages, the overexpressed genes (Figure A6F) were associated with various innate immunity-related processes including type I IFN production, the MDA-5 signalling, the classical pathways of the complement activation, the clearance of apoptotic cells, TLR binding, and the NF-kB transcription factor activity (Figure 8B). In Th cells, the combination treatments led to increases in genes associated with type I IFN production and signalling and MDA-5 signalling in active Th (Figure 8C) and proliferating active Th (Figure 8D), and with the IFN-β production in the proliferating active Th (Figure 8D). In exhausted proliferating Tcyt upregulation of genes associated with the type I interferon (IFN) production, the MDA-5 signalling pathway, the negative regulation of virus replication, the regulation of the macrophage apoptotic process, the regulation of the adaptive immune response and the somatic recombination of immune receptors built from IgG superfamily domains were seen (Figure 8E). Finally, in B cells, positive regulation of type I IFN signalling pathways as well as genes associated with the IgG-mediated immune response and defence responses to virus infections were seen in the combination vs. sham-exposed controls (Figure 8F).

3.6.4. Combination Treatments vs. Reovirus-Alone and BH-Alone Groups

When combination treatments were compared to the reovirus-only treatments, the reovirus-treated intermediate neutrophils had higher levels of genes associated with the TAP2-dependent class I MHC complex (Figure A9A1). In inactive neutrophils, the reovirus treatments increased the expression of genes associated with the negative regulation of type I IFN production, the negative regulation of the adaptive immune responses and IL-2 production (Figure A9A2). In the inactive neutrophil subset, reovirus treatments decreased the expression of genes associated with the positive regulation of the NF-kβ activity and TNF superfamily cytokine production compared to the combination treatments (Figure A9A3). When the combination treatments were compared to the BH-only treatments, differences were seen in macrophages and active Th. In BH-exposed macrophages, downregulated genes were associated with the TLR and NF-kβ signalling, and responses to exogenous dsRNA (Figure A9B1), whereas in active Th, BH treatments upregulated genes that were associated with responses to the macrophage colony-stimulating factor (Figure A9B2). No significant differences were seen in the remaining cell types when the combination treatments were compared to the reovirus-only or the BH-only treatments.

4. Discussion

In this study, murine orthotopic KPC tumours were exposed to either reovirus alone, cavitation-inducing BH alone, or both (combination treatment). The histological appearance of the tumours was studied 6 and 72 h after the end of the BH or sham-BH treatments, and showed the destruction of parts of the tumour after ultrasound treatments compared to the sham-exposed and reovirus-treated tumours (Figure 2). The immune response of the subjects was investigated 24 h after treatment. All KPC tumours including those from control subjects contained immune cells that could create an anti-immune pro-tumour TME, including neutrophils overexpressing genes associated with an N2-like phenotype, CD8⁺ lymphocytes overexpressing immune checkpoint genes, and SIRPα⁺ macrophages (Figure 3 and Table 2 and Table A3). The modulation of the expression of thousands of immune cell genes in treated subjects compared to controls was demonstrated with the combination treatments, more than doubling the number of modulated genes compared to either the reovirus or BH treatments alone or when added together (Table 5). All treatments (reovirus alone, BH alone, and combination) increased the transcription of genes associated with the innate immune response, including type I IFN production and signalling (Figure 7, Figure 8 and Figure A7). In addition, reovirus increased gene expression associated with the viral defence responses (Figure A7), BH upregulated genes associated with TLR3 and NF-kβ signalling, but downregulated genes associated with the TAP complex (Figure 7). The latter was also seen after the combination treatments (Figure 8). The combined treatments increased the abundance of the active and total neutrophils (Figure 4A) and the NLR (Figure A3) in the tumours, which was not seen with the reovirus or BH-only treatments. Finally, ultrasound-driven decreases in the abundance of Treg led to a change in the immune architecture of the pancreatic TME, as demonstrated by the significant increases that were seen in the ratios of multiple cell subsets to Treg (Figure 5).

Extension to current knowledge: The immune effects of “thermal” and “mechanical” focused ultrasound in cancer have been investigated in the past by a number of research groups. Assessments have commonly used antibody-based histological analysis and flow cytometry, with few studies performing holistic transcriptional or proteomics studies. Consequently, little consensus exists about if and how an anti-cancer-specific immune response is activated after ultrasound treatments. Furthermore, some of the effects seen in pre-clinical studies in vivo, such as, for example, the abscopal effect, have not been convincingly replicated in the clinic. This necessitates a more detailed understanding of the immune mechanisms associated with focused ultrasound and immunotherapy treatments. Our study extends the depth of analysis shown in published studies [27] including our own [24]. Previously, the activation of components of the immune system was demonstrated in the Pan02 murine subcutaneous pancreatic cancer model (which carries no KRAS or P53 mutations) using flow cytometry and bulk RNA sequencing (Qiagen “Cancer Inflammation and Immunity Crosstalk” SuperArray platform). Hendricks-Wenger and colleagues reported the upregulation of 8/100 immune genes 24 h after cavitation cloud histotripsy exposures, using a 1 MHz, eight-element histotripsy transducer generating short pulses of <2 cycles, at a prf of 250 Hz and an exposure time of 1 sec, hence sending 250 shots at each treatment point. Upregulated genes included the pro-inflammatory CCR7, CXCL9, CXCR4, GZMB, IL13, IL22, IL5, and HSP90AB1 genes. These initial increases were followed by upregulation of anti-inflammatory genes being seen at the survival endpoints, days later. Of the eight immune cell types investigated using flow cytometry at three different time points (1, 7, and 14 days after treatment), including granulocytes (CD45⁺CD11C⁻Ly6C⁻Ly6G⁺ cells), statistically significant modulation of macrophages, DC, and Treg was seen 14 days after treatment, and of Th 7 days after exposure [27]. In our study, a reduction in the transcription of GZMB in the macrophages and DC in the tumours of BH-treated subjects compared to the sham-exposed controls was seen. The two forms of histotripsy (boiling and cavitation cloud) use the same fundamental mechanism of tissue damage (collapsing bubbles), so these results demonstrate the need for detailed studies to investigate thoroughly their immune mechanisms. Also, we used an orthotopic tumour model based on KPC cells (with G12D KRAS and TP53 mutations) injected into the pancreas of subjects to replicate the native environment in which pancreatic tumours grow. Although the anatomical features differ between humans and mice (e.g., human PDACs typically extend beyond the organ and intertwine with major veins and arteries), the KPC mutations are among the most frequently seen in human PDACs, with KRAS present in >90% and p53 mutations present in >50% of human PDACs. The murine KPC tumours also have a dense collagen stroma [28] and compete with the neighbouring normal pancreas and other organs for nutrients and resources.

Franks and colleagues have published a series of studies on the effects of mechanical therapy ultrasound using a relatively low P- of 6 MPa (below the range accepted as inducing BH, 10–20 MPa [23]) to treat murine breast cancer and melanoma. Bulk proteomic and transcriptomic approaches were used that showed transcriptome and proteome modulation in the ultrasound-treated tumours which depended on the tumour type [29,30,31]. Abe et al. used focused ultrasound at 1.5 MHz, 2% dc, 200 W, and a prf of 5 Hz, for 20 s treatments in a HER2-transduced preclinical murine breast cancer model to induce mechanical damage with or without the addition of anti-PDL1 (given intraperitoneally 3 and 6 days after the ultrasound treatment), and investigated the transcription of genes 8 days after exposures using single-cell transcriptomics [32]. In CD8⁺ T cells, all treatments (ultrasound only, PDL-1 only, and combination) increased the expression of genes associated with type I IFN-mediated signalling, activated T-cell proliferation, and promoted chemokine and cytokine secretion compared to untreated controls. Genes related to CD8⁺ T cell activation including CD33, CX3CR1, CXCL3, CXCL11, and CXCL16 were overexpressed >2-fold in the combination treatment group (ultrasound and anti-PDL1) compared to the group using mechanical ultrasound therapy alone, providing a possible explanation for the enhanced anti-tumour effects that were observed for the combination vs. ultrasound treatment group in the same study. The use of non-transduced “immune cold” pre-clinical cancer models should provide further data about the effects and mechanisms associated with the ultrasound and anti-PDL1 treatments.

BH-induced acoustic cavitation and histological assessments: Acoustic cavitation detection of broadband emissions is generally accepted to indicate the violent collapse of clouds of oscillating bubbles (inertial cavitation), and the release of energy that could lead to tumour disruption. This is a stochastic process, and as such, bubble clouds with unpredictable numbers of bubbles can be created upon ultrasound exposure, and the energy released from each bubble can vary significantly. Tissue architecture in and around the target volume will also affect this process, resulting in some tumour exposure positions having a higher propensity for acoustic cavitation than others. As a result, identical ultrasound exposure pulses are unlikely to produce identical acoustic cavitation outcomes when delivered repeatedly to a single location or different positions within a target tumour. Instead, they can create a range of inertial cavitation activity from undetectable levels, up to levels high enough to saturate detection equipment.

In this study, the number of broadband-positive pulses and the appearance of hyperechoic regions after treatment with BH were used as indicators of the presence of acoustic cavitation activity (both shown in Table 1). Ultrasound treatment of the KPC tumours resulted in detectable broadband signals during the treatment of every tumour. Hyperechoic region formation was also seen on real-time ultrasound images in all tumours except one which had an unusually bright background echogenicity within the tumour before treatment. Histological evidence indicated the destruction of ultrasound-treated tumours by ultrasound-induced acoustic cavitation. A tumour architecture composed of intact, densely packed nuclei was seen within the cells of tumours in the sham and reovirus-treated groups (Figure 2A–C), whereas enlarged intercellular spaces and oddly shaped nuclei were seen in KPC tumours at our first assessment time point 6 h after ultrasound treatments (Figure 2D1,E2). By 72 h, the disrupted zones of the ultrasound-treated tumours had developed into necrotic-like regions with few (or no) intact KPC cells (Figure 2F2,G2), suggesting that extensive BH-induced cell death had taken place in the core of the tumours.

Cluster annotation: PDAC tumours have been considered to be immune “cold” tumours with few immune cells compared to, for example, melanomas. Here we have demonstrated that the murine KPC TME has a diverse array of immune cell infiltrates associated with it (Figure 3 and Figure A2). To understand the role of these infiltrates, they first had to be identified as precisely as possible, and for this reason the UMAP clusters were annotated manually using conserved biomarkers for each immune cell type. Annotation decisions were based on the following hallmarks:

Neutrophils: Seven (7) neutrophil subsets were distinguished (Table 2A,B). They all expressed various protein levels of Ly6G (Figure A2D), overexpressed the genes ITGAM [33], FCGR3 [34], SELL, S100A8, and IFITM1 and were negative for CCL2 differentiating them from macrophages. Also, they expressed the CD14 gene which is expressed by neutrophils under inflammatory conditions [35,36]. All neutrophil subsets showed an N2-like (pro-tumour growth, immune suppressive) gene expression pattern by overexpressing ENTPD1 [37], CXCR2 [33], S100A9 [38], IL10R [39,40], CD177 [41,42], OSM [43], TGFβR [44] (all but the active neutrophils subset cluster), PDL1, and MMP9 [45], and by downregulating pro-inflammatory genes like NOX2 [46], IFN-γ, and IFN-γR that are commonly associated with an N1-like phenotype. All neutrophil clusters overexpressed the activation marker gene CD63 [47], TLRs, and TNF (Table 2A). These seven clusters were further separated into four large and three small clusters. The first of the large clusters was the activated-neutrophil cluster because these cells increased the transcription of genes associated with neutrophil activation including MMP8, a neutrophil degranulation marker [48], several INF-γ-inducible genes including CXCLL10, IFITM6, IFIT2, and TNFAIP6 which have been associated with the activation of innate immune response and M2 macrophage polarisation [49], and MRGPRA2B, the protein product of which acts as a detector of foreign elements on the surface of neutrophils (Table 2B). Another large neutrophil cluster was the intermediate-neutrophil cluster because, here, the transcription of activation genes was lower than that seen in the activated neutrophils cluster (e.g., MMP8, IFN-γ-responsive genes) or not expressed at all (TNFAIP6, MRGPRA2B, CXCL10 [33]) (Table 2B). The third large neutrophil cluster was annotated as the inactive-neutrophils because it showed no overexpression of all the above-mentioned biomarkers (Table 2B). In addition, three smaller clusters were distinguished. The cells in the active-neutrophils subtype cluster expressed similar biomarkers to the active neutrophils including several IFN-stimulated genes (e.g., IFITM6, IFIT2, IFIT3, IFIT1) and MMP8, but did not overexpress CXCL10 and Dach1 (Table 2B). The intermediate-neutrophil subtype cluster had a similar expression pattern to the intermediate neutrophils, and in addition, expressed high levels of the SIGLECF gene (Table 2B). Neutrophils that overexpress SIGLECF have been shown to promote tumour growth and the suppression of the immune system [50]. The inactive-neutrophil subtype had a similar expression pattern to the inactive neutrophils lacking the overexpression of the IFN-stimulated genes, CXCL10, TNFAIP6, and MMP8. Again, this cluster showed high overexpression of SIGLECF and IL23A; the latter was not seen in the inactive-neutrophil cluster (Table 2B). Finally, the transitioning-neutrophil cluster was identified as such due to its location in the UMAP and the overexpression of genes associated with both the active and intermediate neutrophils. Annotation decisions for all remaining clusters are discussed in the manual annotation discussion section of Appendix A.

Collectively, these cluster annotation results demonstrate the existence of a KPC TME that is hostile to the host by favouring tumour growth and the regulation of the anti-cancer immune response in some ways: Firstly, neutrophils, key cells involved in the innate immune response, exhibited N2-like gene expression patterns which are generally associated with pro-tumourigenic processes, whereas N1-like neutrophils would favour a pro-inflammatory phenotype. Secondly, the majority of the Tcyt overexpress exhaustion marker genes such as the immune checkpoints CTLA4, PDCD1, LAG3, TIM, and TIGIT suggestive of the regulation of their cytolytic potential. Thirdly, the presence of a distinct cluster of SIRPα⁺ macrophages suggests that at least some phagocytic cells would be prone to stimulation by tumour-expressed CD47, thereby making them incapable of engulfing abnormal cellular bodies and limiting their anti-tumour role [51,52].

Changes in immune cell abundance: Compared to the sham-exposed group, reovirus treatments decreased the abundance of macrophages (Figure 4B), whereas BH increased transition neutrophils (Figure 4A) and killed Treg (Figure 4C). Combination treatments decreased both the macrophages (Figure 4B) and Treg (Figure 4C) similarly to the reovirus-only and BH-only treatments, respectively. In addition it increased the abundance of active and total neutrophils (Figure 4A) and the NLR which has been postulated as a blood marker of poor prognosis [53] (Figure A3). These results suggest that the combined treatments activate the innate immune response, more than the reovirus and BH treatments when used alone, by increasing the infiltration of the tumours by neutrophils. But, ultimately, this may prove to be a “double-edged sword”. The overexpression of several immune-regulatory genes by these neutrophils suggests that their recruitment to the pancreatic TME may regulate the further activation of the immune response. Experiments designed to block these neutrophils would shed more light on their role in the response of the tumours to the BH and reovirus treatments.

Activated Tcyt and NK cells attack and kill cells that are recognised as foreign, including reovirus-infected and tumour cells. In this study, exposure of the KPC tumours to the combined treatments increased the ratio of the NK cells (Figure 5B), exhausted Tcyt and proliferating Tcyt (Figure 5D) to Treg compared to the sham-exposed controls. It is thus reasonable to suggest that these cells have the potential to induce cell death 24 h after treatment independently of the regulatory functions of Treg. Also, the increases seen in the ratio of other pro-immune cells to Treg such as the DC (Figure 5B), inflammatory monocytes (Figure 5B), and active and proliferating active Th (Figure 5C) are likely to create a more favourable immune TME in the combination-treated tumours compared to the sham-exposed tumours.

Functional enrichment analysis—All treatments: Treatment of the KPC tumours with reovirus or BH resulted in the overexpression of genes associated with the activation of the innate immune response.

Reovirus treatments modulated the transcription of genes in cells of monocytic lineage (macrophages and inflammatory monocytes) whereas only modest DEG was seen in the remaining clusters (Table 5). In macrophages, upregulated genes including SIGLEC1, DHX58, XAF1, IRF7, OAS3, IFIH1, OAS2, TLR3, MORC3, OTUD5, PARP9, STAT2, SP100, USP18, and ADAR (Figure A6A) have been associated with type I IFN signalling and production (Figure A7). Type 1 IFNs include the IFN-α and IFN-β cytokines which play key roles as communication signals for the activation of the innate immune responses against the viral infections. Genes associated with viral defence responses and/or the classical complement activation pathway (Figure A7A1), for example C4B, CLU, RTP4, ZNFX1, DDX60, PML, IFIT3, DTX3L, MX1, EIF2AK2, IRF9, IL15, BST2, MLKL, and IFIT1 were also observed to be overexpressed (Figure A6A), suggesting the central role of these cells in the clearance of the virally infected cells.

BH treatments significantly modulated gene expression in active and intermediate neutrophils, whereas the downregulation of genes was seen in macrophages and DC (Table 5). In active neutrophils, overexpressed genes including IL1RAP, ADAM8, TLR4, CX3CR1, FLOT1, and IL18 (Figure A6B) have been associated with NF-kβ and TLR3 signalling (Figure 7A1,B). NF-kβ is a multifaceted transcription factor that coordinates the inflammatory response and the response of cells to cellular stress. TLR3 signalling is associated with the response of cells to double-strand (ds)RNA. It is reasonable to assume that the BH-destroyed KPC and other stroma cells would be the source of this dsRNA. On the other hand, in macrophages, BH treatments downregulated genes including IL27RA, CD3E, CD96, CD226, CD2, and PGLYRP1, which have been associated with the regulation of type II IFN production, and GZMB, SH2D1A, NKG7, LAG3, CTLA4, TNFSF11, CAMK2N1, ETS1, LILRA5, and IFNG (Figure A6C) that have been associated with the regulation of the lymphocyte and NK cell-mediated immunity (Figure 7C). These processes would be expected to play a role in a cancer-specific adaptive immune response. In DC, downregulated genes including CD160, NKG7, CD8A, ITK, TRBC1, TRBC2, LCK, CD3D, CD3G, CD3E, and CD28 (Figure A6D) have been associated with the T cell and NK cell immune response and the class I MHC (Figure 7D), wheras in active neutrophils BH treatments downregulated genes such as TAP2, PSMB8, H2.Q7, and H2.K1 (Figure A6B), which have been associated with the class I MHC and TAP complexes (Figure 7A2). The class I MHC plays a key role in the presentation of intracellular antigens/peptides to cytotoxic T cells that would then be expected to elicit a sustained adaptive immune response against these antigens. This process is helped by the TAP complex which is responsible for transporting peptides from the cytosol to the cell surface so that they are then presented by the MHC. These results suggest that while BH activated at least some parts of the innate immune response, at the same time it downregulated components associated with the adaptive immune response in various cells. In this context, Price, Bullock and colleagues have recently shown, in a pre-clinical melanoma B16-F10 model using 3 ms BH exposures of P⁻ = 21 MPa, duration of 10 s, prf = 4 Hz, spacing every 1mm, treatment planes separated by 2mm, the release of free antigen into lymph vessels and its acquisition by antigen-presenting cells within 24 h of treatment [54], thus providing at least one way by which the effects of BH on some antigen presentation processes in the pancreatic TME, as seen in this study, could be overcome.

The combination treatments resulted in the modulation of a much higher number of genes compared to the reovirus-only and BH-only treatments in cells of granulocytic (active, intermediate, transitioning, and inactive neutrophils) and monocytic origin (macrophages, inflammatory monocytes, and MDSC-enriched cells) (Table 5). Overexpressed genes in active neutrophils (e.g., S100A9, S100A8, ADAM8, CFLAR, TLR4, CX3CR1, TRAF1, IL1RAP, and FLOT1) (Figure A6E) have been associated with the NF-kβ and TLR3 signalling (Figure 8A1) as was seen for the BH-alone treatments. Overexpressed genes in macrophages (e.g., C4B, SIGLEC1, IRF7, TLR3, DHX58, XAF1, OAS2, DDX60, TLR6, and ZNFX1) (Figure A6F) have been associated with the complement activation and type I IFN signalling and production (Figure 8B), as was seen for the reovirus-only treatments. The combined treatments also increased genes that have been associated with the MDA5 signalling pathway in various cell types (Figure 8) including DDX60, IRF7, DHX58, and OAS3 (Figure A6F). This pathway is an anti-viral immune response mechanism that leads to the production of type I IFNs which would then elicit downstream immune processes. IFN-responsive genes that have been shown to play a role in the host’s anti-viral defences and that were upregulated by the combined treatments in active, intermediate, transitioning and inactive neutrophils included IFITM6 [55], IFI27L2A [56], IL15 [57], and CD14 [58] (Figure 6B). Additionally upregulated genes included the neutrophil and lymphocyte chemoattractants CXCL3 [59], CXCL2 [60], and CCR1 [61], and PLAUR which protects cells against viruses [62] and facilitates cell migration [63] thus showing one mechanism by which neutrophils would have been attracted in the combination-treated KPC TME (Figure 4A). In response to the combination treatments neutrophils also upregulated genes which are associated with overcoming cell stress such as IER-3, an apoptosis inhibitor that protects cells from TNFα induced apoptosis [64] and HILPDA which enables cells to overcome hypoxia [65]. Neutrophil genes associated with a pro-tumour growth phenotype were also upregulated after the combination treatments compared to controls, including the CCL6 gene which promotes immunosuppressive networks, including anti-immune M2 macrophage polarisation [66], and tumour progression [67,68] and LILRB4 that has been associated with negative regulation of the immune response, antigen capture and presentation, and is a potential immunotherapy target [69,70]. These are not unexpected results considering that the neutrophils in the KPC TME express N2-like transcriptomic signatures, as shown earlier in this study (Table 2). The two commonly upregulated genes in the active neutrophils, macrophages, active proliferating Th, and exhausted proliferating Tcyt included the gene XAF1, a tumour suppressor that is upregulated in virus-infected cells and enhances IFN-induced apoptosis [71,72], and PYDC4, which is predicted to regulate the activation of the inflammasomes, which are intracellular bodies that coordinate the response of cells to viral infections [73] (Figure 6C). These results suggest that the combined treatments activate the innate immune response to resolve the reovirus infection and to limit the damage caused to the cellular structures by the BH treatments, but also attract to the tumour immune cells like neutrophils that, in this tumour model, show a pro-tumour N2-like phenotype.

Compared to the sham-exposed controls, the combined treatments, downregulated active neutrophil genes including TAP2, PSMB8, TAPBPL, ATP2A3, HSPD1, ZBTB1, ERAP1, PSME1, and CTSS (Figure A6E) that have been associated with antigen processing and presentation, and the regulation of T cell-mediated immunity (Figure 8A2). Some of these genes were downregulated across neutrophil subsets, for example, PSMB8 and TAPBP (Figure 6A). PSMB8 directs the creation of immunoproteasomes which distinguish and degrade foreign proteins eg, those originating from viruses [74]. TAPBP plays a role in the presentation of cancer cell antigens by antigen-presenting cells [75]. These results suggest that the downregulation of the TAP-dependent class I MHC, at this early time point, is likely an ultrasound-driven event as described above for the BH-only treatments (Figure 7A2). We hypothesise that this could benefit the host by reducing the number of intracellular self-antigens exposed on the class I MHCs of the ultrasound-treated tumours, thereby reducing the risk of an immune response being mounted against self-antigens.

Study limitations: This study has used only one time point for the single-cell immune analysis (24 h). The primary reason for this is that our previous study [24], which investigated the immune response in PDAC at longer time points after treatment (12-day assessment for CD8⁺ and CD4⁺ cells), showed limited beneficial effects on the ultrasound-treated PDAC tumour growth and immune response at that late stage. To understand those diminishing effects, the 24 h time point was used here as a reasonable indicator of the innate immune response activation phase. Another limitation of this study is that the KPC cells were not investigated. This is because there is a trade-off between the depth of analysis possible with current single-cell technologies, cost, and the number of samples assessed. To achieve unparalleled sequencing depth, we chose to concentrate our investigation on the repertoire of the tumour-associated immune cells and to investigate each tumour as a unique sample, thereby providing a better illustration of the variability seen between and within treatment groups. Finally, in future studies, transcriptomic results will need to be complemented by protein and enzyme activity data to fully validate the various cell subtypes and their regulation seen in this study.

5. Conclusions

In this study, the acute immune response of the orthotopic murine pancreatic KPC tumours 24 h after treatment with reovirus, cavitation-inducing BH (boiling histotripsy), and the combination of BH and reovirus was compared to both sham-exposed controls and each other. Our data shows that (a) the pancreatic TME consists of immune cell subsets programmed to create a hostile environment for the host TME by overexpressing immune signatures that are associated with an N2-like phenotype in neutrophils, an exhaustion phenotype in cytotoxic lymphocytes, and an inactivation phenotype in a macrophage subset; (b) all treatments overexpressed genes associated with the activation of the innate immune response compared to the sham-exposed controls, with the effects being more marked after the combined treatments which showed higher levels of tumour-infiltrating neutrophils and higher modulation of gene expression in the immune cells than the reovirus and BH treatments used alone; (c) the ultrasound-based treatments (BH and combination) downregulated genes associated with antigen presentation, possibly as a mechanism to limit activation of the immune system against self-antigens; and (d) BH and combination treatments changed the balance of pro-immune cells to Treg in the pancreatic TME, thereby creating a more favourable pro-immune tumour micromilieu. These results demonstrate that the combined BH and reovirus treatment enhances the immune response compared to the treatments used alone. They also show some of the potential mechanisms by which the immune TME can facilitate the immune evasion capabilities of pancreatic tumours. In future studies, the BH and reovirus treatments will be combined with immunotherapies (e.g., neutrophil depletion) to investigate the effects of these cells on pancreatic tumour growth.