Patients were vaccinated with 200 µg of ArgLong2 (ARG1_169-206), a 38-aa peptide (ISAKDIVYIGLRDVDPGEHYILKTLGIKYFSMTEVDRL), and 100 µg of PD-L1Long1 (PD-L1_19-27), a 19-aa peptide (FMTYWHLLNAFTVTVPKDL). The peptides were provided by Polypeptide (Strasbourg, France).

The two vaccines were administered at the same time. Briefly, the peptides were individually dissolved in 500 µg sterile water/DMSO and emulsified with 500 µl Montanide ISA-51, just prior to administration. The vaccines were administered subcutaneously, one in each shoulder, every two weeks. Patients received six treatments (i.e., 12 vaccinations) over 12 weeks. After these first six treatments, treatment was paused for 1-3 months; then, six more treatments were given, again at two-week intervals. An additional six treatments could be scheduled for patients that showed a response. The entire treatment plan is shown in Supplementary Figure 2 .

Adverse events were assessed according to the Common Terminology Criteria for Adverse Events, version 5.0. All patients were evaluated prior to inclusion with a medical examination by the treating physician. This included spleen palpation (no ultasonography nor computed tomography was performed), blood sample analyses (Hemoglobin, leukocyte differentiation count, platelets, IgG, IgA, IgM, Hematocrit, Bilirubin, Potassium, Sodium, Creatinine, albumin, uric acid, lactate dehydrogenase, Alkaline Phosphatase, Alanine transaminase, amylase, bilirubin, D-dimer, ionized calcium, C-Reactive Protein, Thyrotropin, thyroxin, Luteinizing Hormone, Adrenocorticotropic Hormone, Cortisol, Hepatitis B, hepatitis C (IgG), HIV, HTLV-1(IgG), IgG and IgM for Cytomegalovirus (CMV), Epstein-Barr Virus (EBV) and toxoplasmosis) and an electrocardiogram. According to the treatment plan, patients were evaluated at inclusion, during treatment pause, and at the end of the trial (EOT). At every vaccine treatment, an investigator recorded the patient’s symptoms, and when necessary, conducted a medical examination and evaluation. Bone marrow biopsies were acquired from patients before trial entry and at end of the trial. Histopathological evaluation of nonblinded biopsies was performed by a trained hematopathologist.

We analyzed patient DNA samples for mutations with the droplet digital PCR (ddPCR) method, on a QX100 system (Bio-Rad, Hercules, California, USA), according to manufacturer instructions. Briefly, we mixed 10 µl of 2x digital PCR Supermix for probes (Bio-rad), 2 µl primer/probe mix, 3 µl nuclease-free water, and 5 µl DNA (20 ng/µl), for a total volume of 20 µl. Droplets were generated on a QX100 Droplet Generator System (Bio-rad). The polymerase chain reaction (PCR) was performed with an initial stage at 95°C for 10 min, followed by 43 cycles of 94°C for 30 sec and 57°C for 60 sec, then a final stage at 98°C for 10 min. PCR was carried out on an Applied Biosystems Veriti 96-Well Thermal Cycler (Thermo Fisher, Waltham, MA, USA). Droplets were subsequently quantified on a QX100 Droplet Reader (Bio-rad) and analyzed with Quantasoft™ Analysis Pro software.

The JAK2V617F primer/probe assay included a forward primer: GCTTTCTCACAAGCATTTG, a reverse primer: GCATTAGAAAGCCTGTAGTTTTA, and two probes: Fam-TCGTCTCCACAGAaACATACTCCATGAGACGA-BHQ1 (mutant c.1849G>T) and Hex-TCGTCTCCACAGACACATACTCCATGAGACGA-BHQ1 (wildtype).

Next generation sequencing was performed using the Illumina Ampliseq Myeloid panel including 40 genes (ABL1, ASXL1, BCOR, BRAF, CALR, CBL, CEBPA, CSF3R, DNMT3A, ETV6, EZH2, FLT3, GATA2, HRAS, IDH1, IDH2, IKZF1, JAK2, KIT, KRAS, MPL, MYD88, NF1, NPM1, NRAS, PHF6, PRPF8, PTPN11, RB1, RUNX1, SETBP1, SF3B1, SH2B3, SRSF2, STAG2, TET2, TP53, U2AF1, ZRSR2, WT1). Genomic DNA was purified from peripheral blood at baseline and 9 months after the first vaccination. Libraries were prepared using the Ampliseq for Illumina Myeloid Panel protocol, and 2 × 150 bp paired-end sequencing was done on the NextSeq 500 platform (Illumina^® Inc, San Diego, CA, USA). The Illumina Sequencing Analysis Viewer (SAV) software was used for quality control of the sequencing runs. Alignment of sequencing data to the human reference genome (GRCh37/hg19) and variant calling of mapped reads were performed in CLC Genomics Workbench software v.22. The VarSeq™ software v.2.2.4 (Golden Helix, Inc., Bozeman, MT, USA) was applied for annotation and filtering of variants. Variants with coverage <100x, a variant allele frequency (VAF) <1%, and germline, introns, and SNPs with minor allele frequency >1% (ExAC variant frequencies, Broad Institute, MA, USA) were excluded from further analysis.

For peripheral blood mononuclear cell (PBMC) isolation, blood samples were obtained and cryopreserved as previously reported (31) at baseline, after three vaccinations, after six vaccinations (during the treatment pause), after seven vaccinations, after nine vaccinations, and at EOT. Heparinized bone marrow aspirations (10 mL in a heparinized tube) were obtained at baseline and at the end of the trial. Ortho-Lysing Buffer diluted 10× in H₂0 was added to half of the sample, followed by centrifugation and incubation for 15 minutes in the dark. The other half of the sample was handled and cryopreserved following the same procedure as for PBMCs.

To assess the presence of PD-L1 and ARG1-specific lymphocytes, we performed a delayed-type hypersensitivity (DTH) test. DTH tests were assessed at baseline and at EOT. Briefly, we administered intradermal injections of the two peptides, without the adjuvant, at the lower back. The peptides were dissolved in sterile water and DMSO. We also administered a control injection, which included water and DMSO, but no peptide. At 48 h after the injection, skin reactions (induration) were measured; additionally, at the sites of ArgLong2- and PD-L1long1 injections, punch biopsies were acquired and cut into fragments. To identify fragments that contained skin-infiltrating lymphocytes (SKILs), fragments were cultured in 24-well plates in RPMI-1640 with 10% human serum and 100 U/mL interleukin-2 (IL-2), with penicillin, streptomycin, and fungizone, for 3–5 weeks to allow SKIL outgrowth. Every second or third day, half the medium was replaced with fresh medium containing IL-2, penicillin, streptomycin, and fungizone. After 3–5 weeks, SKILs were harvested, and a fraction was tested in ELISPOT assays. The remaining SKILs were cryopreserved.

Immune responses were evaluated with in vitro and ex vivo IFN-γ ELISPOT assays. For in vitro ELISPOT assays, PBMCs were thawed and stimulated with the target epitope. The next day, PBMCs were stimulated with IL-2 (120 U/mL) and incubated for 12-14 days. The PBMCs were then counted and plated in ELISPOT wells. Cells were restimulated with or without the target peptide. All conditions were performed in triplicates. For ex vivo IFN-γ ELISPOT assays, cells were thawed, rested, then plated on ELISPOT plates. Cells were stimulated with or without the target epitope for 24-48 h to ensure antigen presentation.

In vitro ELISPOT assays were performed with a cell density of 2.5 ×10⁵ cells/well. Ex vivo ELISPOTs were performed with cell densities of 9 ×10⁵ cells/well, for PBMCs, and 6.8×10⁵ cells/well, for bone marrow mononuclear cells. Plates were analyzed with the ImmunoSpot Series 2.0 Analyzer (CTL, Shaker Heights, Ohio). Results were generated by subtracting the background obtained with negative controls. A detailed description of our setup was described previously (32). Statistical significance of ELISPOT responses was analyzed by the DFR method (33). For non-triplicate samples, responses were evaluated empirically and defined as true if the number of spots observed in the peptide stimulated wells were at least double of the spot counts in the control wells.

To evaluate the phenotypes of T cells that responded to stimulation in IFN-γ ELISPOT assays, we conducted intracellular cytokine staining (ICS). The in vitro stimulation was similar to that described above in the IFN-γ ELLISPOT section. Briefly, after 12-14 days in vitro culture, PBMCs were restimulated with peptide. After 1 h, Brefeldin A was added to inhibit protein transport. After 4 additional hours of incubation, the cells were stained with T-cell surface markers CD4-PerCP (cat. 345770), CD8- FITC (cat. 345772), CD3-APC-H7 (cat. 560275) and a dead cell marker FVS510 (564406) (all from BD Biosciences). Samples were then fixed and permeabilized using eBioscience™ Fixation/Permeabilization buffers (eBioscience, cat. 00-5123-43, 00-5223-56) and stained with IFNg-APC (cat.341117, BD Biosciences), TNFa-BV421 (cat.562783, BD Biosciences) in eBioscience permeabilization buffer (eBioscience, cat. 00-8333-56) and analyzed on FACSCanto™ II (BD Biosciences) using BD FACSDiva software version 8.0.2 as described previously (32).

Analysis was performed on samples derived from patient PBMCs and bone marrow-derived MNCs isolated at baseline and end-of-trial. For the PBMCs, CD14 Microbeads (Miltenyi Biotec) were utilized to separate CD14+ and CD14- cells. Total RNA purification was performed using the RNEasy Plus Mini Kit (Qiagen) according to the manufacturer’s protocol. RNA concentration was measured on a NanoDrop2000 Spectrophotometer (Thermo Scientific). cDNA was synthesized using the iScript™ cDNA Synthesis Kit (Bio-Rad) based on 500 ng or 400 ng total RNA (PBMC-derived cells and bone marrow, respectively). RT-qPCR analysis was performed using the TaqMan Gene Expression Assay on a Roche LightCycler 480 instrument. The assay was performed in technical triplicates for all primers and the subsequent data analysis was performed as previously described using the ΔΔCt-method (34) with normalization of PD-L1 (Primer ID Hs01125296_m1) and ARG1 (Primer ID Hs00163660_m1) expression to the housekeeping gene POL2RA (Primed ID Hs00172187_m1) and to the baseline sample. Controls lacking reverse transcriptase during cDNA creation were included for primer validation.

PBMCs collected at three time-points and bone marrow collected at 2 time-points were thawed and washed in preheated phosphate-buffered saline. Fc-receptors were blocked by incubating with human IgG (50 μg/ml), and dead cells were stained with the Fixable Near-IR Dead cell stain kit (Thermo-Fisher). After mixing, the cells were stained in the dark at 4°C for 20 min with fluorochrome-labeled antibodies ( Supplementary Figure 3 ). Next, the cells were washed and analyzed with a NovoCyte Quanteon Flow Cytometer (Agilent, Santa Clara, CA). The gating strategy is described in Supplementary Figure 4 . Data were analyzed with NovoExpress 1.5.1 software. All gates at baseline were applied to all timepoints. Illustrations were created with Graphpad Prism v 8.0 (GraphPad Software. Inc.).

ELISPOT responses were analyzed with the distribution free resampling (DFR) method (33). DFR analyses were performed with the R statistical analysis program. Immune subsets at different time points were compared with the Wilcoxon matched-pairs signed-rank test. P values ≤0.05 were considered significant. Immune subset analyses were performed in Graphpad Prism v 8.0 (GraphPad Software. Inc.).

We evaluated bone marrow biopsies and blood samples from 12 patients before inclusion in the vaccination trial. Three patients did not meet the inclusion criteria and were excluded: One patient had patient had unmeasurable JAK2V617F and normal bone marrow and thus did not meet the diagnostic criteria, one patient had progressive disease and proceeded to alloHSC and one patient had additional findings in the bone marrow suggesting a more MDS-like disease and thereby did not meet the diagnostic criteria Thus, nine patients (5 male, 4 female, median age: 57 years, range: 46 to 72) with a median disease duration (time from diagnosis) of 2 years (range: 4 months to 10 years) were enrolled in the clinical study. The vaccines were administered over a period of six to nine months. Among the nine patients, eight received a minimum of 12 vaccines; one patient did not receive one ArgLong2 vaccine. One patient (#3) proceeded with an extra round of vaccinations, due to an apparent drop in the JAK2V617F allele burden, measured with qPCR. This drop could however not be confirmed by ddPCR. The primary end of trial (EOT) was defined as 12 treatments.

At inclusion, all nine patients had a measurable JAK2V617F mutation burden, ranging from 1.36% to 32.58% (median 12.81%). Additionally, two patients harbored bystander mutations in the DNMT3A gene as determined by next generation sequencing. Seven patients were diagnosed with ET, and two patients were diagnosed with PV (29). At the time of inclusion seven out of nine patients received cytoreductive treatments for MPN: these treatments included IFN-α (N=5, Pegasys, 4 patients received 45 ugx1sc/week, 1 patient received 135 ug x 1 sc/every fourth week), ANA (N=1), and phlebotomy (N=1). The median platelet count for the patient cohort was 489 × 10⁹/L (range: 219 – 630 × 10⁹/L). The median hemoglobin count was 8.6 mmol/L (range: 7.8 – 9.7 mmol/L); the median hematocrit was 0.41 (range: 0.39 – 0.46) and the median leucocyte count was 5.2 × 10⁹/L (range: 3.5 – 8.2 × 10⁹/L). Patient baseline characteristics are summarized in Table 1 . At the primary EOT, the JAK2V617F mutation burden ranged from 1.33% to 37.25% (median 14.64%; Figure 1A ). Bone marrow samples and blood samples were evaluated to investigate any potential impact of the vaccines on hematological response and molecular response in addition to changes in disease phenotype. However, we did not identify any signs of neither a molecular ( Figure 1A ) nor a hematological response on average ( Figures 1 B–D , Supplementary Figure 5 ). No alterations in bone marrow architecture and cellular composition were identified in any of the patients (data not shown).

Vaccines were generally well tolerated: the majority of reported adverse events were grades 1 and 2 ( Table 2 ). Injection site reaction (grade 2) was the most commonly observed adverse event and was reported by all the patients at least once during the study period. In two patients, injection site reactions remained visible for one year after the last vaccine (patients #4 and #5). Only one grade 3 adverse event was registered: one patient (# 9) had a vasovagal reaction immediately after the first PD-L1Long1 vaccination and did not receive the ArgLong2 peptide vaccine during the same hospital visit. However, the patient continued the treatments afterwards and received both vaccines, as scheduled, without similar reactions. There were two peculiar adverse events reported by the same patient during the trial: a change in taste (grade 1 dysgeusia) and a change of body odor (grade 1). The patient reported on this at the time of the last vaccine, but claimed that the changes started already after the first vaccine. The reactions were deemed to be related to the DMSO contained in the vaccine solution. Patient # 1 reported several reactivations of Herpes Simplex virus during the trial (at treatments 3, 5, 7, and 11). The patient had a history of Herpes Simplex infections, but the last reactivation had occurred several years before study inclusion.

To assess the vaccine induced immune responses, we analyzed PBMCs and BMNCs for T-cell responses against the two peptides used in the trial by ex vivo IFN-γ ELISPOT assay ( Figure 2 ). We detected a spontaneous response against the ARG1- and the PD-L1-derived peptides in PBMCs from one out of nine and from three out of nine patients at baseline respectively. During treatment, the responses against the ARG1-derived and PD-L1-derived peptides were detected ex vivo in PBMCs from five and eight patients respectively ( Figures 2A–D ). In patient 5, we did not observe a significant ex vivo PBMC response during treatment, even though the patient’s PBMCs displayed a spontaneous T-cell response at baseline. Ex vivo IFN-γ ELISPOT was performed using bone marrow mononuclear cells (BMNC) from patients 4 and 6, and cells from both patients showed vaccine-specific T cells after treatment with an apparent increase in responses in EOT samples compared to baseline ( Figures 2E, F ). Due to low viability of the isolated BMNC, we were not able to perform a similar assay on BMNCs from the remaining seven patients.

In an addition to ex vivo assays, we performed an in vitro IFN-γ ELISPOT and intracellular cytokine staining (ICS) on in vitro pre-stimulated PBMCs to improve the detection of the treatment induced response and phenotypically characterize the vaccine reactive T cell populations ( Figure 3 ; Supplementary Figures 6 D–F ). At baseline, we detected spontaneous responses against the ARG1-derived peptide in the PBMCs of six of nine patients, albeit most of the responses were low in magnitude similarly with the ex vivo ELISPOT results. During the vaccine trial, all patients but patient 2, demonstrated an enhanced ARG1-specific T cell response with a clear increase in ArgLong2 response magnitude already after 3^rd vaccination ( Figure 3A ; Supplementary Figures 6A, C ). PBMCs from six out of nine patients displayed a spontaneous response against the PD-L1-derived peptide at baseline, with a majority of the baseline responses being low in magnitude ( Figure 3B and Supplementary Figure 6B ). Only PBMCs from patient 8 did not display a spontaneous response against neither the PD-L1 nor ARG1 epitopes at baseline. PBMCs from all patients displayed greatly enhanced T-cell responses against the PD-L1-derived epitope after 3 vaccinations ( Figure 3B ; Supplementary Figures 6B, C ).

In ICS we chose to analyze PBMCs collected at time points in which the individual patients showed the strongest immune response as detected by in vitro ELISPOT ( Figures 3A, B ). We detected that the majority of the vaccine specific T cells were producing TNF-α and/or IFN-γ in response to peptides used in the vaccination. Responses against ArgLong2 were identified in all nine patients, though primarily in the CD4⁺ T-cell compartment with lower incidence of CD8⁺ T-cell responses detected in 6 out of 9 patients ( Figures 3C, E ). Similarly, the analyses confirmed that all 9 patients generated CD4⁺ T-cell response against PD-L1-derived peptide and 6 out of 9 patients also displayed a lower magnitude of CD8⁺ T cell response against the same peptide ( Figures 3D, F ).

Induction and expansion of vaccine specific T cell responses was also evaluated by delayed type hypersensitivity (DTH) testing. Six out of nine patients provided informed consent for this analysis. Baseline DTH biopsies collected from four out of six patients could not be expanded to obtain skin infiltrating lymphocytes (SKILs) cultures. In baseline SKILs samples obtained from patients 2 and 9 we could only detect a significant response against the ARG1-derived peptide in IFNγ ELISPOT in patient 9 ( Supplementary Table 1 ). At EOT we expanded SKILs from biopsies collected from all six patients and in 5 out 6 analyzed patients, the SKILS exhibited vaccine-specific responses against one or both peptides as determined by IFNγ ELISPOT signifying expansion and target epitope dependent homing of the vaccine induced T cells ( Supplementary Table 1 ). None of the patients displayed a visible skin reaction at baseline, however at EOT (corresponding to the expanded SKILs samples), all patients except patient 1 showed skin reactions with redness and swelling to various degrees at all injected DTH sites.

We investigated the peripheral blood of the patients for changes in composition of immune cell subsets during the vaccination trial. We performed fluorescence-activated cell sorting (FACS) on isolated PBMC at three time-points: baseline, treatment pause (after 6 treatments), and EOT. Bone marrow aspirations acquired at baseline and at EOT were also analyzed.

The fraction of CD3⁺ T cells in PBMCs remained unchanged throughout the study period ( Figure 4A ). However, the percentage of CD4⁺ T cells decreased significantly from baseline to EOT, and accordingly the fraction of CD8⁺ T cells increased significantly during treatment ( Figures 4B, C ). The significant increase in the percentage of CD8⁺ T cells was detectable already after six treatments. We further analyzed the changes in the CD8⁺ T cell subsets and observed a significant decrease and increase in the CD8⁺ T cell T_EM and T_EMRA subpopulations respectively by the end of the treatment ( Figures 4F, G ). A significant decrease in T_EM cells was already detectable at the treatment pause. T_EMRA cells increased from baseline to EOT. Additionally, while no changes were seen in the T_CM population, a significant decrease in CD8⁺ T_Naïve cells was seen ( Figures 4D, E ). The CD4⁺ T-cell subsets remained unchanged, including central memory (T_CM), effector memory (T_EM), T_Naive, and T_EMRA cells ( Supplementary Figure 7 ). PD-1 expression on CD3⁺ T cells as well as CD4⁺ and CD8⁺ T cell subsets remained unchanged ( Supplementary Figures 8A–C ). No changes in levels of regulatory T cells, natural killer (NK) cells, or B cells were observed during treatment (data not shown). CD16^dimCD56^hi NK cells decreased significantly during the trial. MDSCs, defined as HLA^-DR^-CD33⁺CD14⁺ cells, monocytes, and dendritic cells remained stable during treatment. CD3⁺-, CD4⁺-, CD8⁺- and NK-cells in the bone marrow remained stable during treatment (data not shown).

Changes in the expression of PD-L1 were assessed by flow cytometry and no alterations of PD-L1 expression on the myeloid cells (CD3^-CD19^-CD56^- cells) were observed during the trial ( Supplementary Figure 8D ). Changes in PD-L1 and ARG1 expression were also evaluated in PBMCs (n=9) and BMNCs (n=6) of the treated patients using RT-qPCR. Interestingly, a significant decrease in the PD-L1 mRNA expression was seen in CD14⁺ myeloid cells in the peripheral blood from baseline to end-of-treatment ( Figure 4H ). Decrease in PD-L1 expression was also detected in CD14^Neg cell fraction in 6 out of 7 evaluated patient samples ( Supplementary Figure 8E ). Interestingly, while no ARG1 expression was detected in the peripheral blood in neither CD14⁺ nor CD14^Neg cell fractions, a clear decrease in the ARG1 expression was detected in BMNCs of 5 out of 6 evaluated patients ( Figure 4I ). Contrary to ARG1, expression of PD-L1 was only detected in BMNCs of one patient (#5) and an increase in PD-L1 expression was seen in this sample ( Supplementary Figure 8F ). These results suggest a potential differential expression of ARG1 and PD-L1 between the peripheral blood and the bone marrow compartments in patients with MPN.