Nivolumab and Ipilimumab in Metastatic Melanoma Are Associated with Distinct Immune Landscape Changes and Response-Associated Immunophenotypes

The reshaping of the immune landscape by nivolumab (NIVO) and ipilimumab (IPI) and its relation to patient outcomes is not well-described. We used high-parameter flow cytometry and a novel computational platform, CytoBrute, to define immunophenotypes of up to 15 markers to assess peripheral blood samples from metastatic melanoma patients receiving sequential NIVO>IPI or IPI>NIVO. The two treatments were associated with distinct immunophenotypic changes and had differing profiles associated with response. Only two immunophenotypes were shared but had opposing relationships to response/survival. To understand the impact of sequential treatment on response/survival, phenotypes that changed after the initial treatment and differentiated response in the other cohort were identified. Immunophenotypic changes occurring post-NIVO were predominately associated with response to IPI>NIVO, but changes occurring post-IPI were predominately associated with progression with NIVO>IPI. Among these changes, CD4+CD38+CD39+CD127-GARP- T-cell subsets were increased after IPI treatment and were negatively associated with response/survival for the NIVO>IPI cohort.

• Jeffrey S. Weber: oversaw project, conceived experiments, interpreted data, edited manuscript. Mirati Therapeutics, Iovance Biotherapeutics, Cue Biopharma, Fate Therapeutics, Atra 29 Biotherapeutics, and Fortress Biotech. 30 • Andressa S. Laino: None. 31 (clusters 1, 6, and 7) and CD73 (cluster 1), T-cells with a naïve-like phenotype (cluster 2), as well as 1 cell types that are not well-defined by the markers we measured (clusters 3-5 and 8), Eight clusters 2 were also identified for the immunophenotypes decreasing post-NIVO, as shown in Supplemental 3 Figure 3C. The phenotypic composition of these clusters is shown in Supplemental Figure 3D. 4 Cell types that decrease with NIVO were predominately characterized expression of markers 5 associated with central memory T-cells (e.g. CD45RO+, CCR7+, CD127+, CD95+). 6 In IPI-treated patients, 4,498 immunophenotypes were increased and 2,679 were reduced 7 relative to baseline with a p-value <0.05 ( Figure 2B). For immunophenotypes increasing post-IPI, 8 eight clusters were identified (Supplemental Figure 4A); these were defined by the markers listed 9 in Supplemental Figure 4B. In contrast to NIVO-associated changes, many of these phenotypes 10 were characterized by CD45RO+ and CD95+. Nine clusters were identified from the 11 immunophenotypes decreasing post-IPI (Supplemental Figure 4C), for which the phenotypic 12 breakdown is shown in Supplemental Figure 4D. 13

NIVO and IPI have distinct impacts on the peripheral immunophenotypic landscape 15
To better compare the immunophenotypic effects of the two drugs, we examined which 16 changes were common to both drugs versus unique to only one drug. For the two treatments, 584 17 immunophenotypes that changed overlapped (~5%), as shown in Figure 2C. Figure 2D shows that 18 of the 584 immunophenotypic changes common across drugs, 281 increased and 244 decreased. 19 However, 56 immunophenotypes had reciprocal changes; these cell populations were expanded 20 post-IPI but contracted post-NIVO. Only three phenotypes were increased post-NIVO but decreased 21 post-IPI. Supplemental Figure 5A shows the immunophenotypes that increased with both 22 treatments fell into seven clusters. A CD4+CD38+ T-cell phenotype predominated in clusters 2, 3, 23 4, and 7 (Supplemental Figure 5B). The 244 immunophenotypes that decreased with IPI or NIVO 24 grouped into seven clusters (Supplemental Figure 5C), which included almost exclusively CD4+ 1 immunophenotypes with the exception of cluster 7 (Supplemental Figure 5D). Six clusters 2 represented the immunophenotypes decreasing post-NIVO but increasing post-IPI (Supplemental 3 Figure 5E). Clusters 1 and 4 were composed of CD4+OX40+Ki67+ T-cells (Supplemental Figure  4 5F). 5 We next evaluated whether changes in circulating immunophenotypes could distinguish 6 treatment regimens. To do so, the difference in frequency between week 13 and week 0 (delta value) 7 was calculated for each immunophenotype. These delta values were then fed into an EN regularized 8 regression model with repeated cross validation. Based on paired changes of the 584 identified 9 overlapping phenotypes from all patient samples, the model was able to predict whether paired 10 patient samples were from those who received NIVO or IPI with an area under the curve (AUC) of 11 0.867. The corresponding receiver operating characteristic (ROC) is shown by the black dotted line 12 in the left panel of Figure 2E. The model values for individual paired samples are shown in the right 13 panel of Figure 2E. Patient outcomes were added as a color dimension with responding patients 14 (partial or complete response according to RECIST 1.1 criteria) in blue and progressing patients in 15 red. Using the EN model determined by all patient samples, we then determined the ROC for 16 responding and progressing patients separately. As shown in the left panel of Figure 2E, an AUC of 17 0.982 was achieved in responding patients (blue line) and an AUC of 0.714 in progressing patients, 18 suggesting that patients who respond to therapy have more distinct immune changes than non-19 responders. 20 21 Peripheral blood immunophenotypes at baseline and post-treatment are associated with patient 22 outcomes after NIVO>IPI sequential therapy. 23 We next sought to determine if baseline and/or week 13 (post-first agent) peripheral blood 1 immunophenotypes were associated with patient outcomes. In baseline samples from NIVO>IPI 2 treated patients (left panel, Figure 3A), 260 signatures were associated with both response to 3 therapy (p<0.05, Mann-Whitney U-test comparing responders and progressors) and overall survival 4 (divided above and below median frequency, p<0.05, Mantel-Cox test). Each dot represents a 5 significant immunophenotype and is colored by the associated p-value from the comparison of 6 frequency differences between responders and progressors. The x-coordinate is the median 7 frequency of the immunophenotype in progressors and the y-coordinate is the corresponding median 8 frequency in responders. The 61 immunophenotypes significantly elevated at baseline in responding 9 patients formed five clusters, as shown in Supplemental Figure 6A. Cluster 1 consisted of 10

T-cells. 17
Post-NIVO (week 13), 662 immunophenotypes were found to be associated with patient 18 outcomes in those receiving sequential NIVO>IPI treatment (right panel, Figure 3A). In contrast to 19 baseline response-associated immunophenotypes which were predominately CD4+, nearly all the 20 post-NIVO response-associated immunophenotypes were CD8+, in particular those found at higher 21 frequencies in responders. Amongst these immunophenotypes, 561 were increased in frequency in 22 responding relative to progressing patients. As shown in Supplemental Figure 6E, ten clusters were 23 formed from these immunophenotypes. Combinations of CCR7, CD127, CD45RO and CD95 24 positive CD8+ T-cells comprised Clusters 2, 3, 4, 5, 6, 8 and 10. A total of 101 immunophenotypes 1 forming seven clusters were found to be decreased in responding relative to progressing patients 2 (Figure 3A right panel and Supplemental Figure 6G). As shown in Supplemental Figure 6H, 3 Clusters 1, 2, and 6 were composed of CD8+CD45RA+ T-cells. Collectively, a number of 4 immunophenotypes that were associated with patient response and survival and were grouped into 5 several immunophenotypically-related clusters were identified. These included elevated levels of 6 naïve-like (CD45RA+, CD127+) T-cells and CD8+LAG3+ phenotypes at baseline and elevated 7 levels of central-memory-like (CD45RO+, CCR7+, CD127+, CD95+) T-cells post-NIVO. A population 8 of CD4+CD38+CD39+ T-cells at baseline was also associated with progression and shorter survival. 9 10 Peripheral blood immunophenotypes at baseline and post-treatment are associated with patient 11 outcomes after IPI>NIVO therapy. 12 In IPI>NIVO treated patients, 432 baseline immunophenotypes were associated with 13 response to treatment and survival ( Figure 3B, left panel). Of these, 376 were elevated in frequency 14 in responding relative to progressing patients. These 376 immunophenotypes formed eight clusters 15 and were a mixture of CD4+ and CD8+ T-cell populations (Supplemental Figures 7A and 6B). 16 Clusters 5 and 6 were composed of CD4+ T-cells expressing LAG3+ and GARP+. 56 phenotypes, 17 forming five clusters, were decreased in frequency in responding patients (Supplemental Figure  18 7C and 7D). All clusters contained a predominance of CD45RO+ and CD95+ immunophenotypes, 19 while Clusters 2, 4 and 5 also contained CCR7-expressing immunophenotypes. 20 At week 13, after IPI treatment ( Figure 3B, right panel), 668 immunophenotypes were 21 associated with treatment response and survival, with 608 elevated and 60 at lower frequencies in 22 responding compared to progressing patients ( Figure 3B, right panel). As shown in Supplemental 23 Figure 7E, immunophenotypes higher in relative frequency in responders formed seven clusters. 24 CD4+ and CD8+ T-cell immunophenotypes with CCR7+ expression comprised Clusters 1, 2, 3, 5 1 and 7 (Supplemental Figure 7F). The 60 immunophenotypes with lower frequencies in responders 2 formed 6 clusters (Supplemental Figure 7G). These clusters were mixed populations of CD4+ and 3 Figure 7H). Cluster 2 included T-cells expressing CD38+ and CD39+. 4 5 Distinct immunophenotypes are associated with patient response in NIVO>IPI and IPI>NIVO treated 6 patients. 7

CD8+ T-cells (Supplemental
We next compared the immunophenotypes associated with outcome between NIVO>IPI and 8 IPI>NIVO patients. As shown in Figure 3C, only two (related) immunophenotypes were associated 9 with response in both cohorts. The associations were reciprocal between the two cohorts. These 10 cells were CD14+CD11C+CD33+CD15-CD19-CD66B-PDL1-PDL2+CD163+GAL9-CD80-CD86-11 41BBL+CD40+OX40L+ and an identical immunophenotype in which CD66B was not measured. The 12 frequency of this immunophenotype in responding and progressing patients for each treatment 13 cohort is shown in Figure 3D. In NIVO>IPI treated patients, higher frequencies of these 14 immunophenotypes were associated with progression (p=0.0164), while in IPI>NIVO treated 15 patients, higher frequencies were associated with patient response (p=0.0246). This was reflected 16 in the survival curves shown in Figure 3E. Frequencies of these phenotypes above the median were 17 associated with shorter survival in NIVO>IPI treated patients (p=0.0050, HR 95% CI: 1.681-18.74) 18 but with prolonged survival in IPI>NIVO treated patients (p=0.0463, HR 95% CI: 1.019-9.038). 19 20 NIVO-associated immune landscape changes favor response-associated immunophenotypes in the 21

IPI>NIVO cohort. 22
In the Checkmate 064 trial, the NIVO>IPI treatment arm had greater rates of response and 23 overall survival compared to the IPI>NIVO treatment arm 4 . We hypothesized that the 24 immunophenotypic impact of treatment with NIVO or IPI may have altered the immune landscape in 1 a manner that influenced subsequent response or resistance to the second agent. We tested this 2 hypothesis by determining the overlap between the immunophenotypes that changed significantly 3 after the first treatment in the regimen and those associated with treatment response and survival in 4 the opposing cohort. As depicted in Figure 4A, of the 3,959 cell populations that were significantly 5 altered after NIVO, four overlapped with those that were associated at baseline with response and 6 survival in IPI>NIVO treated patients. However, none of the markers we measured were expressed 7 by the cells in these four immunophenotypes (i.e. no positive markers) (data not shown). An 8 additional 95 overlapping immunophenotypes associated with treatment response/patient survival at 9 week 13 in the IPI>NIVO cohort and changed after NIVO treatment were found. Ninety-nine percent 10 (94/95) of these overlapping immunophenotypes were positively associated with response, with 11 almost all (93/94) immunophenotypes increasing post-NIVO and were elevated at week 13 in 12 responders to IPI>NIVO. 13 We next evaluated the composition of the 95 immunophenotypes identified above associated 14 with treatment response/patient survival at week 13 in the IPI>NIVO cohort (Supplemental Figure  15 8A). The 93 immunophenotypes that were increased at week 13 post-NIVO and associated at 16 baseline with response in IPI>NIVO (green dots in Supplemental Figure 8A) are described in 17 Supplemental Figure 8B. These cells were all CD4+ and predominately CD45RO-and CCR7+. A 18 representative population of CD4+CD45RO-CCR7+ T-cells is shown in Figure 5A. The leftmost 19 panel shows that most patients had an increase in the frequency of this population post-NIVO 20 treatment (p=0.035). The second panel shows that higher frequencies of this population in IPI>NIVO 21 treated patients at week 13 were associated with response (p=0.0046). The second to right panel 22 shows that IPI>NIVO patients with greater than the median frequency of CD4+CD45RO-CCR7+ T-23 cells at week 13 have longer overall survival (p=0.019, HR 95% CI: 1.343-27.42). To validate this 24 result, we evaluated changes in the frequency of this immunophenotype in a separate cohort of 1 metastatic melanoma patients treated with nivolumab monotherapy (ClinicalTrials.gov identifier 2 NCT01176461). Forty paired patient samples were assessed. The rightmost panel of Figure 5A  In a similar assessment of the overlap between post-IPI changes and NIVO>IPI outcome-9 associated immunophenotypes, we identified 74 immunophenotypes that were altered by IPI 10 treatment and whose frequency at baseline in the NIVO>IPI treatment arm was associated with 11 response/survival ( Figure 4B). Similarly, we identified 36 immunophenotypes that were altered by 12 IPI, and whose presence at week 13 in the NIVO/IPI treatment arm was associated with 13 response/survival ( Figure 4B). Supplemental Figure 8C shows how these immunophenotypes 14 cluster into groups of cell populations. 15 Unlike the NIVO-associated changes, IPI-associated changes had a negative association with 16 NIVO response. At baseline, 92% of the overlapping phenotypes (68/74; Figure 4B) were associated 17 with progression; 67 of these immunophenotypes were increased post-IPI and found at relatively 18 lower frequencies in NIVO>IPI responding patients. As shown in Supplemental Figure 8D, these 19 immunophenotypes were all CD4+CD38+CD39+GARP-CD127-and varied with respect to other 20 markers.
A representative population from these 67 immunophenotypes, 21 CD4+CD38+CD39+CD127-HELIOS-CD25-LAG3-CXCR3-CCR4-LAP-GARP-T-cells, is shown in 22 Figure 5B. The leftmost panel shows that this population was increased post-IPI (p=0.039). The 23 second panel shows that significantly higher frequencies of this population are seen in progressing 24 patients in the NIVO>IPI cohort (p=0.0079). The second to right panel shows that patients with 1 greater than median frequency of this population also have significantly shorter survival (p<0.0001, 2 HR 95% CI: 4.399-69.23). To independently validate the association of this immunophenotype with 3 patient response, we assessed the frequency of CD4+CD38+CD39+CD127-GARP-T-cells in 4 peripheral blood samples of metastatic melanoma patients treated with nivolumab monotherapy. 5 Twenty responding and 47 progressing patient samples were assessed. Shown in the rightmost 6 panel of Figure 5B, elevated levels of that immunophenotype were associated with progression of 7 disease (p=0.047). 8 At week 13, all (36/36) of the phenotypes altered by IPI were associated with poor treatment 9 response/survival for NIVO>IPI patients. Of the 36 immunophenotypes identified, IPI increased the 10 frequency of 30 of these cell populations, but patients responding to NIVO>IPI had significantly 11 decreased frequencies of these cells. Supplemental Figure 8F shows how these 12 immunophenotypes clustered and Supplemental Figure 8H describes the marker composition of 13 these clusters. The additional 6 overlapping immunophenotypes were decreased post-IPI but had 14 elevated frequencies at week 13 in NIVO>IPI responding patients. These six immunophenotypes 15 included both CD4+ and CD8+ T-cells and were predominately CD127+, CD95+, CD45RO+ and 16 CCR7+. A representative phenotype from these six is shown in Figure 5C. The leftmost panel shows 17 that this population was significantly decreased post-IPI (p=0.011). The middle panel shows that the 18 frequency of this population at week 13 in NIVO>IPI patients is significantly higher in responders 19 compared to progressors (p=0.035). The second to right panel shows the associated increased 20 survival in patients with >median frequency of this population (p=0.016, HR 95% CI: 1.322-15.49). 21 However, shown in the rightmost panel of Figure 5C, in a cohort of metastatic melanoma patients 22 treated with NIVO, we were unable to confirm a relationship between reduced levels of these cells 23 and disease progression (p=0.530).

Elastic Net model. 3
To determine if the changes in overlapping immunophenotypes described in Figure 4A and 4 4B were sufficient to predict patient outcomes, we used patient delta values for the 209 overlapping 5 immunophenotypes and an EN model with LOO cross validation. As shown in Figure 4C, responding 6 and progressing patients in the NIVO>IPI-treated cohort were accurately categorized with an AUC 7 of 0.918. Figure 4D shows that patients in the IPI>NIVO treated cohort were correctly categorized 8 as responders or progressors, but with lesser sensitivity and specificity, resulting in an AUC of 0.786. 9 These results further support the importance of immunophenotypic changes resulting from treatment 10 and their relation to patient response. 11 12

Discussion. 13
In the current study we utilized a novel and powerful approach to analyzing high-dimension 14 flow cytometry data to assess the impact of the immune checkpoint inhibitors nivolumab and 15 ipilimumab on the peripheral blood immune landscape. By assessing the frequencies of complex 16 immunophenotypes in lieu of dimension-reduction analytical methods (e.g. tSNE), we were able to 17 more precisely identify treatment-associated changes in the immunophenotypic landscape, identify 18 response-associated immunophenotypes, and assess their relationships. Collectively, these data 19 suggest that IPI and NIVO alter the peripheral blood immunophenotypic landscape of patients in 20 distinct ways. Further, IPI-associated alterations overlapped with immunophenotypes associated 21 with progression of disease and shorter survival in the NIVO>IPI cohort. Although the overlap 22 occurred with the patient responses to sequential NIVO>IPI noted at week 25, overall responses 23 deviated little from week 13 responses (post-NIVO alone) 4 . These results suggest that IPI-associated 24 immune landscape changes may impair responses to subsequent NIVO and may explain the lower 1 response rate and shorter survival seen in the IPI>NIVO cohort. 2 While the emergence of single-cell, high-dimension technologies has increased the ability to 3 probe the anti-tumor immune response, approaches to analyze the complex data generated have 4 failed to keep pace. Dimension-reduction techniques such as principal component analysis and 5 tSNE are commonly used to visualize high parameter data, but these techniques do not report the 6 specific combination of markers expressed by cell types. A critical, unanswered question in high-7 parameter data analysis is whether clustering and dimension-reduction algorithms sufficiently 8 capture all the cell types that differ across study groups. Classically, to exhaustively examine cell 9 phenotypes in lower parameter flow cytometry datasets, analysis has been performed by using 10 combinatorics, a method that constructs all possible phenotypes from the markers measured. For 11 example, using combinatorics, an n-color flow cytometry experiment would report the number of cells 12 expressing each combination of expression (+) or lack of expression (-) for the n markers, resulting 13 in 2 n phenotypes. The use of a neutral condition for each marker allows assessment of shorter and 14 simpler phenotypes, resulting in 3 n phenotypes. The identification of simple phenotypes is critical for 15 generating translatable discoveries from high parameter technology. After all, it would be 16 unnecessarily complex and expensive to develop clinical tests to detect cell populations that are 17 defined by much more than three markers, when simpler surrogates might readily be identified within 18 a high parameter dataset. 19 While combinatorics offers a means to precisely identify and quantify all cell subsets in a 20 sample, it is computationally intensive. Even for a 10-parameter experiment measuring one million 21 cells, combinatoric analysis using the R-based flowType algorithm requires four hours; in 22 comparison, the same analysis is completed by our CytoBrute platform within two seconds. The 23 improvement in computational time is attributable to the unique and proprietary distributive 24 computing approach CytoBrute uses, and this approach is adaptable to other R-based algorithms 1 and applications, including machine learning-based analysis of high parameter datasets. 2 There are shortcomings in this study that impact the interpretation of the data. Rather than 3 using traditional FDR approaches, we utilized a non-multiple comparison adjusted p-value of <0.05 4 as a determination of significance. This approach was taken for several reasons. First, the number 5 of samples available for assessment were limited, and in turn the lower threshold of p-values 6 obtainable was limited. Second, the combinatoric nature of the CytoBrute approach creates non-7 independent measurements which would be highly overcorrected if using traditional multiple 8 comparison corrections. However, given 1) the use of leave one out cross validation in EN models, 9 2) the validation of important identified immunophenotypes in an independent cohort, and 3) our 10 focus on demonstrating that the impact of IPI and NIVO are distinct and that the immune landscape 11 changes induced by IPI are associated with lower response to NIVO and shorter survival in the 12 IPI>NIVO cohort, the conclusions of this study are supported by the data. Many response-associated 13 immunophenotypes not overlapping with treatment changes were also identified that have not yet 14 been evaluated in a validation cohort. While beyond the scope of this study, these represent 15 potentially important biomarkers and are the subject of future validation efforts. 16 Several thousand significant changes in peripheral blood immunophenotypes were observed 17 after NIVO or IPI treatment, highlighting the systemic impact of these agents. These changes were 18 largely distinct with only <5% overlap in significantly changed immunophenotypes. Taken with data 19 presented showing a lack of overlap in outcome-associated immunophenotypes between the two 20 cohorts/treatments and published literature 6,7,8,9 it is increasingly clear that the systemic immune 21 impact and the mechanisms by which CTLA4 and PD1 blockade function are distinct. 22 Of the few overlapping immunophenotypes, both treatments increased populations of 23 CD4+CD38+CD39+CD127-GARP-T-cells, which were found to be associated with poor outcomes. 24 This population was also associated with progression in an independent cohort of NIVO-treated 1 patients, validating the importance of this phenotype in patient response. To the authors' knowledge, 2 this population of T-cells has not been previously described. CD39 is an ectonucleotidase that 3 converts extracellular ATP into adenosine and has been the focus of many studies demonstrating 4 its roles in generating an immunosuppressive resulting in its emergence as a potential therapeutic 5 target 17,18,19,20,21,22 . CD38 also functions as an ectoenzyme, both as a hydrolase and converts 6 NAD+ to cyclic ADP-ribose. It is being assessed as a target in combination with immunotherapy 23, 7 24 . In agreement with our data, CD38 expression was recently shown to be upregulated post-8 immunotherapy and was found to be associated with negative outcomes in a murine checkpoint 9 inhibition model and in human patients 25 . Based on the current knowledge of the function of CD38 10 and CD39 and our observation that CD4+CD38+CD39+ immunophenotypic clusters are associated 11 with poor patient outcomes, we hypothesize that this population is immunosuppressive. 12 Investigations of mechanistic relationship to patient response, the function of this population and the 13 efficacy of targeting it are warranted by the data presented in this study and are ongoing. 14 In addition to the low level of overlap of immunophenotypic changes, no immunophenotypes 15 associated with patient outcomes were found to be shared between the treatment sequences, with 16 the exception of two related immunophenotypes. This further highlights the distinct impact of the two 17 therapies. The two overlapping immunophenotypes were reciprocally associated with patient 18 outcomes in the two cohorts. Based on the expression of CD14, CD11b and CD33, these cells are 19 likely of myeloid origin 26 and expressed a mixed inflammatory (e.g. 41BBL+, CD40+) 27, 28 and 20 suppressive phenotype (e.g. PDL2+, CD80-, CD86-) 29, 30, 31 . While beyond the scope of the present 21 study, this cell population will be further interrogated in functional assays to determine potential 22 mechanisms by which it could be associated with the disparate outcomes in the two treatment 23

cohorts. 24
T-cell phenotypes associated with memory subsets were consistently and significantly 1 associated with patient outcomes in both treatment cohorts. In the NIVO>IPI cohort, responding 2 patients had higher frequencies of T-cells with a naïve phenotype (CD4+CD45RA+CD127+) at 3 baseline and central memory phenotypes (CD4/8+CD45RO+CD127+CD95+CCR7+) post-NIVO. 4 Conversely, at baseline, progressing patients had more differentiated effector T-cells 5 (CD4+CD45RO+CD95+) and higher frequencies of CD8+CD45RA+ T-cells post-NIVO. These data 6 suggest that the formation of memory T-cells may be important for the efficacy of NIVO. In contrast, 7 in the IPI>NIVO cohort, responding patients had increased frequencies of CD4/8+CD45RO+CD95+ 8 and CD4+CD45RA+CD95+ T-cells at baseline and increased frequencies of CD4+CD45RO-CCR7+ 9 T-cells post-IPI. Several other immunophenotypic clusters were found to be associated with patient 10 response and survival including CD4+CD45RA+CD127+, CD4+CD45RO+CD95+ and CD8+LAG3+ 11 T cell populations. 12 Collectively, these data suggest that the impact of IPI and NIVO on the immunophenotypic 13 landscape of patients is distinct and that the impact of IPI may be associated with resistance to 14 subsequent NIVO therapy, consistent with poor outcomes in the IPI>NIVO cohort of Checkmate-15 064. In further support of this interpretation, in clinical trials the response rates to NIVO in patients 16 progressing after IPI are lower than those in IPI-naïve patients 32,33,34,35,36 . However, these response 17 rates are compared across different studies and lower response rates to NIVO in IPI-refractory 18 patients may result from selection of immunotherapy-resistant patients. Response rates in the 19 NIVO>IPI cohort were similar to concurrent NIVO and IPI, suggesting that concomitant IPI does not 20 negatively impact NIVO efficacy. Regardless, the data presented herein raise concerns about the 21 negative impact of IPI treatment prior to NIVO and warrant consideration in patient treatment 22 decisions. Future studies will investigate immunophenotypic changes occurring with concurrent 23 treatment and associated patient outcomes. Further, studies will need to address whether the 24 immune landscape changes resulting from IPI are normalized over time and the duration that takes 1 to occur. Future investigations will also need to investigate the function of, and potential value as 2 biomarkers of immune cell populations that are associated with treatment outcomes in this study. 3 4

Acknowledgements. 5
We extended our appreciation to Bristol-Myers Squibb, including Christine Horak and Megan Wind-6 Rotolo, for all of their assistance and feedback in this study.   represents an individual patient sample. (E) A survival plot for this immunophenotype is also shown. 4 Patients were stratified based on median frequency of the immunophenotype. 5 Nivolumab>ipilimumab treated patients with greater than median frequencies are shown in blue and 6 less than median frequency in green. Ipilimumab>nivolumab treated patients with greater than 7 median frequencies are shown in red and less than median are shown in purple. Immunophenotypes significantly increased post nivolumab (p<0.05, Wilcoxon signed-rank test) were 20 reduced to a two-dimensional graph by Uniform Manifold Approximation and Projection (UMAP) by 21 using frequency values for all patient samples as n-dimension variables. Clusters were determined 22 by K-Means. Each dot represents a significant immunophenotype and is colored by the p-value. (B) 23 The frequency for the top markers appearing in each cluster are graphed. Positive markers (e.g. Immunophenotypes significantly increased post ipilimumab (p<0.05, Wilcoxon signed-rank test) 31 were reduced to a two-dimensional graph by UMAP by using frequency values for all patient 32 samples as n-dimension variables. Clusters were determined by K-Means. Each dot represents a 33 significant immunophenotype and is colored by the p-value. The frequency for all markers appearing in immunophenotypes increasing post-ipilimumab and lower 1 in NIVO>IPI responding patients are plotted. (E) The frequency for all markers appearing in 2 immunophenotypes increasing post-ipilimumab and higher in NIVO>IPI responding patients are 3 plotted. (F) Immunophenotypes significantly changed post-ipilimumab and associated with 4 response/survival in NIVO>IPI treated patients at week 13 are plotted as in A and C. (G) The 5 frequency for all markers appearing in immunophenotypes decreased post-ipilimumab and higher in 6 NIVO>IPI responding patients are plotted. (H) The frequency for all markers appearing in 7 immunophenotypes increasing post-ipilimumab and lower in NIVO>IPI responding patients are 8 plotted.