Predicting Cytotoxic T-cell Age from Multivariate Analysis of Static and Dynamic Biomarkers

Abstract

Adoptive T-cell transfer therapy relies upon in vitro expansion of autologous cytotoxic T cells that are capable of tumor recognition. The success of this cell-based therapy depends on the specificity and responsiveness of the T cell clones before transfer. During ex vivo expansion, CD8+ T cells present signs of replicative senescence and loss of function. The transfer of nonresponsive senescent T cells is a major bottleneck for the success of adoptive T-cell transfer therapy. Quantitative methods for assessing cellular age and responsiveness will facilitate the development of appropriate cell expansion and selection protocols. Although several biomarkers of lymphocyte senescence have been identified, these proteins in isolation are not sufficient to determine the age-dependent responsiveness of T cells. We have developed a multivariate model capable of extracting combinations of markers that are the most informative to predict cellular age. To acquire signaling information with high temporal resolution, we designed a microfluidic chip enabling parallel lysis and fixation of stimulated cell samples on-chip. The acquisition of 25 static biomarkers and 48 dynamic signaling measurements at different days in culture, integrating single-cell and population based information, allowed the multivariate regression model to accurately predict CD8+ T-cell age. From surface marker expression and early phosphorylation events following T-cell receptor stimulation, the model successfully predicts days in culture and number of population doublings with R2 = 0.91 and 0.98, respectively. Furthermore, we found that impairment of early signaling events following T cell receptor stimulation because of long term culture allows prediction of costimulatory molecules CD28 and CD27 expression levels and the number of population divisions in culture from a limited subset of signaling proteins. The multivariate analysis highlights the information content of both averaged biomarker values and heterogeneity metrics for prediction of cellular age within a T cell population. Adoptive T-cell transfer therapy relies upon in vitro expansion of autologous cytotoxic T cells that are capable of tumor recognition. The success of this cell-based therapy depends on the specificity and responsiveness of the T cell clones before transfer. During ex vivo expansion, CD8+ T cells present signs of replicative senescence and loss of function. The transfer of nonresponsive senescent T cells is a major bottleneck for the success of adoptive T-cell transfer therapy. Quantitative methods for assessing cellular age and responsiveness will facilitate the development of appropriate cell expansion and selection protocols. Although several biomarkers of lymphocyte senescence have been identified, these proteins in isolation are not sufficient to determine the age-dependent responsiveness of T cells. We have developed a multivariate model capable of extracting combinations of markers that are the most informative to predict cellular age. To acquire signaling information with high temporal resolution, we designed a microfluidic chip enabling parallel lysis and fixation of stimulated cell samples on-chip. The acquisition of 25 static biomarkers and 48 dynamic signaling measurements at different days in culture, integrating single-cell and population based information, allowed the multivariate regression model to accurately predict CD8+ T-cell age. From surface marker expression and early phosphorylation events following T-cell receptor stimulation, the model successfully predicts days in culture and number of population doublings with R2 = 0.91 and 0.98, respectively. Furthermore, we found that impairment of early signaling events following T cell receptor stimulation because of long term culture allows prediction of costimulatory molecules CD28 and CD27 expression levels and the number of population divisions in culture from a limited subset of signaling proteins. The multivariate analysis highlights the information content of both averaged biomarker values and heterogeneity metrics for prediction of cellular age within a T cell population. Immune cell-based therapies hold promise in cancer therapy by harnessing the body's natural defense mechanisms against tumors, while leaving healthy cells unharmed (1.Gattinoni L. Powell Jr., D.J. Rosenberg Adoptive for on Adoptive cell a to cancer adoptive transfer of T cells in for and of adoptive T-cell therapy T cells for the of with of lymphocyte and ex vivo Adoptive transfer of receptor T lymphocyte clones in with Although cancer cells are the is to and tumor cells. Adoptive therapy with cytotoxic CD8+ T cells T of T of relies on the isolation of and tumor T cells and in vitro in the cancer to tumor specificity and in vivo to in vivo and tumor vivo is a for of cells. The of adoptive transfer therapy by the of T cells. from cancer have a cells that expansion against in of T lymphocyte replicative senescence in The of T-cell replicative senescence in of the a for the tumor cells a ex vivo culture T have a ex vivo and expansion in T cell and replicative senescence a limited in vitro To sufficient number of cells before tumor are and several of in the of and to a of senescence by Although specificity for the tumor Adoptive T cell therapy T cell clones for the of with vivo and of T present in and and of the a for are in cell to with and to to in the cancer these replicative senescent cells will not to tumor cells and the of these therapies Powell Jr., D.J. Rosenberg of tumor in adoptive with L. Rosenberg of in vitro the in vivo of T success of adoptive transfer is to T-cell clones before transfer based upon specificity of in vitro expansion of T-cell clones in the of with of and is the loss of expression of costimulatory markers CD27 that cells are by and of and CD28 in in and in T-cell a for both and have been with senescent CD8+ cell biomarkers are not sufficient to accurately the of senescent nonresponsive cells. and that of appropriate to D.J. and in the for biomarkers of D.J. Powell expression in culture of T-cell of in and of expression in D.J. Powell of T-cell clones in culture a of and to and have not to determine the most informative metrics of cellular senescence for the of of T-cell function. To of the of T cells in culture combinations of we a regression analysis from with in vitro expansion to adoptive transfer in The multivariate model developed in the present the of T-cell the of markers and signaling with a microfluidic that samples phosphorylation events by and of stimulated cells. of the and in and to with at the a number of cells and the in a and parallel cell stimulation and lysis for early signaling events in T cell in microfluidic cell and analysis content cell in a microfluidic measurements with cells by and within cells. The signaling cellular and surface expression levels the of the chip allowed the model to determine of to the of replicative senescence The multivariate analysis highlights the of both averaged values and heterogeneity in the cell population for prediction of the replicative senescence within a T-cell of based adoptive T-cell transfer therapy is on the of T-cell clones Powell Jr., D.J. Rosenberg of tumor in adoptive with been T-cell expansion methods in the T-cell have we the culture stimulation have been developed vivo expansion for tumor T cells for adoptive T cell is to CD28 for cytotoxic for adoptive and expansion of cytotoxic T cell clones for adoptive and been a of for tumor a of for tumor To the of impairment of and in senescent cytotoxic T surface marker the of CD28 loss the of markers of replicative senescence and of CD8+ T static expression of surface markers are not for prediction of cell Although CD28 is a biomarker of expression by of T lymphocyte replicative senescence of of CD28 expression by by in cells CD28 loss in senescent T by that cells are to and of and and the loss of CD28 is not with is loss of the CD28 and that of CD28 in and not the number of receptor in to stimulation for cells Jr., of in the of and T with CD28 a marker of and combinations of biomarkers to to predict the of senescence in a population. To T-cell different to early phosphorylation within a following and vivo analysis of T-cell Although are informative in assessing cell on population in and culture of cells in vivo by T The analysis of phosphorylation following T-cell receptor stimulation a for cell and on signaling is for and by cytotoxic T to a information the heterogeneity of the early phosphorylation events in T-cell of measurements at Although a of the acquisition methods the that signaling information is in prediction of cellular age and costimulatory surface marker the is not surface markers information on cellular and predict cell predict phosphorylation a of cell not of of biomarkers and is by the development of and several to model and cellular age in cellular model have been developed of of senescent cells in and and in Quantitative of markers for cell a model that expression with model for age-dependent expression of the tumor analysis of T cells in vitro to senescence analysis to biomarkers D.J. Powell of T-cell clones in culture a of of these on static information from the cells of is for the biomarkers of in vitro for CD8+ T we to a multivariate surface of CD8+ and this we developed a microfluidic and a model to cell responsiveness and accurately predict cell and respectively. been that T cells are to to within a of and temporal of T cell receptor signaling with a The of the microfluidic of the phosphorylation in the following stimulation with and allowed to measurements in a and in at number not cell stimulation and lysis for early signaling events in T cell The to stimulated cells with the of in the regression the microfluidic in the information content from the signaling Although have of to informative in analysis of and by and to L. Quantitative combinations of predict model of signaling a for analysis of by regression of dynamic multivariate these on signaling that for the a number of cells with high temporal and and in we measurements for analysis from of cells not from to within a stimulation is a in biomarkers of the the that by the of signaling of have developed a microfluidic to of signaling to with with on chip on cells content cell in a microfluidic to that and cellular on single-cell on to cell and to cell population signaling with is limited by the number of proteins that the microfluidic chip the and of cell that have the allowed to of while to for the heterogeneity in phosphorylation in with in for signaling in different not with developed regression to of the and to predict age of ex vivo cells. have been to signaling and of the of and by and to L. Quantitative combinations of predict model of signaling a for analysis of by regression of dynamic multivariate of this of that information content is within phosphorylation events in T cells L. Quantitative combinations of predict is to multivariate analysis and in and to and surface From the CD28 and CD27 expression markers in cellular age and to the model not the surface markers and and age in culture, to a of replicative senescence and of CD8+ T because of the culture to been that the and stimulation cells is to CD28 for cytotoxic for adoptive is a marker with the T-cell expression of been on T cells. the of stimulation, in the of cells capable of that of replicative senescence and of CD8+ T have loss of CD28 expression of CD28 expression on T cells is by receptor and and to cell before this model is limited in to the culture a that for expansion model the of early signaling following to and predict cellular age. have in early signaling events in in T from and in T from Jr., in the of proteins to in T the ex vivo culture expansion and major and from the model in accurately cellular age. is the to following to phosphorylation of the and phosphorylation of and T cell in T from lymphocyte signaling and and in in and and Jr., to are for the of have with age for the in in expression in the that this is and depends on the in We that the heterogeneity in expression and cellular is with cellular age. age and of phosphorylation and of a multivariate analysis this a high in to with and in T cell from heterogeneity in multivariate model is to the age and the of a cell population for cells stimulation with multivariate signaling information with cellular model of signaling a for L. Quantitative combinations of predict of a on cell the of to the of the this is not a cell that we and the biomarkers are not in a to the age of the cells. of a a to the the model on the most informative markers of the for and predicts cellular age from We this model to a to age of the of a microfluidic chip the analysis of T cell by in the days and We of and a chip to the early phosphorylation events with in to informative in the regression prediction of population and days in to a cell that cells for of population The fixation and facilitate the of single-cell information with Immune cell-based therapies hold promise in cancer therapy by harnessing the body's natural defense mechanisms against tumors, while leaving healthy cells unharmed (1.Gattinoni L. Powell Jr., D.J. Rosenberg Adoptive for on Adoptive cell a to cancer adoptive transfer of T cells in for and of adoptive T-cell therapy T cells for the of with of lymphocyte and ex vivo Adoptive transfer of receptor T lymphocyte clones in with Although cancer cells are the is to and tumor cells. Adoptive therapy with cytotoxic CD8+ T cells T of T of relies on the isolation of and tumor T cells and in vitro in the cancer to tumor specificity and in vivo to in vivo and tumor vivo is a for of cells. The of adoptive transfer therapy by the of T cells. from cancer have a cells that expansion against in of T lymphocyte replicative senescence in The of T-cell replicative senescence in of the a for the tumor cells a ex vivo culture T have a ex vivo and expansion in T cell and replicative senescence a limited in vitro To sufficient number of cells before tumor are and several of in the of and to a of senescence by Although specificity for the tumor Adoptive T cell therapy T cell clones for the of with vivo and of T present in and and of the a for are in cell to with and to to in the cancer these replicative senescent cells will not to tumor cells and the of these therapies Powell Jr., D.J. Rosenberg of tumor in adoptive with L. Rosenberg of in vitro the in vivo of T cytotoxic T cell of regression T-cell receptor cytotoxic T cell of regression T-cell receptor To success of adoptive transfer is to T-cell clones before transfer based upon specificity of in vitro expansion of T-cell clones in the of with of and is the loss of expression of costimulatory markers CD27 that cells are by and of and CD28 in in and in T-cell a for both and have been with senescent CD8+ cell biomarkers are not sufficient to accurately the of senescent nonresponsive cells. and that of appropriate to D.J. and in the for biomarkers of D.J. Powell expression in culture of T-cell of in and of expression in D.J. Powell of T-cell clones in culture a of and to and have not to determine the most informative metrics of cellular senescence for the of of T-cell function. To of the of T cells in culture combinations of we a regression analysis from with in vitro expansion to adoptive transfer in The multivariate model developed in the present the of T-cell the of markers and signaling with a microfluidic that samples phosphorylation events by and of stimulated cells. of the and in and to with at the a number of cells and the in a and parallel cell stimulation and lysis for early signaling events in T cell in microfluidic cell and analysis content cell in a microfluidic measurements with cells by and within cells. The signaling cellular and surface expression levels the of the chip allowed the model to determine of to the of replicative senescence The multivariate analysis highlights the of both averaged values and heterogeneity in the cell population for prediction of the replicative senescence within a T-cell population. of based adoptive T-cell transfer therapy is on the of T-cell clones Powell Jr., D.J. Rosenberg of tumor in adoptive with been T-cell expansion methods in the T-cell have we the culture stimulation have been developed vivo expansion for tumor T cells for adoptive T cell is to CD28 for cytotoxic for adoptive and expansion of cytotoxic T cell clones for adoptive and been a of for tumor a of for tumor To the of impairment of and in senescent cytotoxic T surface marker the of CD28 loss the of markers of replicative senescence and of CD8+ T static expression of surface markers are not for prediction of cell Although CD28 is a biomarker of expression by of T lymphocyte replicative senescence of of CD28 expression by by in cells CD28 loss in senescent T by that cells are to and of and and the loss of CD28 is not with is loss of the CD28 and that of CD28 in and not the number of receptor in to stimulation for cells Jr., of in the of and T with CD28 a marker of and combinations of biomarkers to to predict the of senescence in a population. To T-cell different to early phosphorylation within a following and vivo analysis of T-cell Although are informative in assessing cell on population in and culture of cells in vivo by T The analysis of phosphorylation following T-cell receptor stimulation a for cell and on signaling is for and by cytotoxic T to a information the heterogeneity of the early phosphorylation events in T-cell of measurements at Although a of the acquisition methods the that signaling information is in prediction of cellular age and costimulatory surface marker the is not surface markers information on cellular and predict cell predict phosphorylation a of cell not of of biomarkers and is by the development of and several to model and cellular age in cellular model have been developed of of senescent cells in and and in Quantitative of markers for cell a model that expression with model for age-dependent expression of the tumor analysis of T cells in vitro to senescence analysis to biomarkers D.J. Powell of T-cell clones in culture a of of these on static information from the cells of is for the biomarkers of in vitro for CD8+ T we to a multivariate surface of CD8+ and this we developed a microfluidic and a model to cell responsiveness and accurately predict cell and respectively. been that T cells are to to within a of and temporal of T cell receptor signaling with a The of the microfluidic of the phosphorylation in the following stimulation with and allowed to measurements in a and in at number not cell stimulation and lysis for early signaling events in T cell The to stimulated cells with the of in the regression the microfluidic in the information content from the signaling Although have of to informative in analysis of and by and to L. Quantitative combinations of predict model of signaling a for analysis of by regression of dynamic multivariate these on signaling that for the a number of cells with high temporal and and in we measurements for analysis from of cells not from to within a stimulation is a in biomarkers of the the that by the of signaling of have developed a microfluidic to of signaling to with with on chip on cells content cell in a microfluidic to that and cellular on single-cell on to cell and to cell population signaling with is limited by the number of proteins that the microfluidic chip the and of cell that have the allowed to of while to for the heterogeneity in phosphorylation in with in for signaling in different not with developed regression to of the and to predict age of ex vivo cells. have been to signaling and of the of and by and to L. Quantitative combinations of predict model of signaling a for analysis of by regression of dynamic multivariate of this of that information content is within phosphorylation events in T cells L. Quantitative combinations of predict is to multivariate analysis and in and to and surface From the CD28 and CD27 expression markers in cellular age and to the model not the surface markers and and age in culture, to a of replicative senescence and of CD8+ T because of the culture to been that the and stimulation cells is to CD28 for cytotoxic for adoptive is a marker with the T-cell expression of been on T cells. the of stimulation, in the of cells capable of that of replicative senescence and of CD8+ T have loss of CD28 expression of CD28 expression on T cells is by receptor and and to cell before this model is limited in to the culture a that for expansion model the of early signaling following to and predict cellular age. have in early signaling events in in T from and in T from Jr., in the of proteins to in T the ex vivo culture expansion and major and from the model in accurately cellular age. is the to following to phosphorylation of the and phosphorylation of and T cell in T from lymphocyte signaling and and in in and and Jr., to are for the of have with age for the in in expression in the that this is and depends on the in We that the heterogeneity in expression and cellular is with cellular age. age and of phosphorylation and of a multivariate analysis this a high in to with and in T cell from heterogeneity in multivariate model is to the age and the of a cell population for cells stimulation with multivariate signaling information with cellular model of signaling a for L. Quantitative combinations of predict of a on cell the of to the of the this is not a cell that we and the biomarkers are not in a to the age of the cells. of a a to the the model on the most informative markers of the for and predicts cellular age from We this model to a to age of the of a microfluidic chip the analysis of T cell by in the days and We of and a chip to the early phosphorylation events with in to informative in the regression prediction of population and days in to a cell that cells for of population The fixation and facilitate the of single-cell information with The of based adoptive T-cell transfer therapy is on the of T-cell clones Powell Jr., D.J. Rosenberg of tumor in adoptive with been T-cell expansion methods in the T-cell have we the culture stimulation have been developed vivo expansion for tumor T cells for adoptive T cell is to CD28 for cytotoxic for adoptive and expansion of cytotoxic T cell clones for adoptive and been a of for tumor a of for tumor To the of impairment of and in senescent cytotoxic T surface marker the of CD28 loss the of markers of replicative senescence and of CD8+ T static expression of surface markers are not for prediction of cell Although CD28 is a biomarker of expression by of T lymphocyte replicative senescence of of CD28 expression by by in cells CD28 loss in senescent T by that cells are to and of and and the loss of CD28 is not with is loss of the CD28 and that of CD28 in and not the number of receptor in to stimulation for cells Jr., of in the of and T with CD28 a marker of and combinations of biomarkers to to predict the of senescence in a population. To T-cell different to early phosphorylation within a following and vivo analysis of T-cell Although are informative in assessing cell on population in and culture of cells in vivo by T The analysis of phosphorylation following T-cell receptor stimulation a for cell and on signaling is for and by cytotoxic T to a information the heterogeneity of the early phosphorylation events in T-cell of measurements at Although a of the acquisition methods the that signaling information is in prediction of cellular age and costimulatory surface marker the is not surface markers information on cellular and predict cell predict phosphorylation a of cell not of of biomarkers and is by the development of and several to model and cellular age in cellular model have been developed of of senescent cells in and and in Quantitative of markers for cell a model that expression with model for age-dependent expression of the tumor analysis of T cells in vitro to senescence analysis to biomarkers D.J. Powell of T-cell clones in culture a of of these on static information from the cells of is for the biomarkers of in vitro for CD8+ T we to a multivariate surface of CD8+ and this we developed a microfluidic and a model to cell responsiveness and accurately predict cell and respectively. been that T cells are to to within a of and temporal of T cell receptor signaling with a The of the microfluidic of the phosphorylation in the following stimulation with and allowed to measurements in a and in at number not cell stimulation and lysis for early signaling events in T cell The to stimulated cells with the of in the regression the microfluidic in the information content from the signaling Although have of to informative in analysis of and by and to L. Quantitative combinations of predict model of signaling a for analysis of by regression of dynamic multivariate these on signaling that for the a number of cells with high temporal and and in we measurements for analysis from of cells not from to within a stimulation is a in biomarkers of the the that by the of signaling of have developed a microfluidic to of signaling to with with on chip on cells content cell in a microfluidic to that and cellular on single-cell on to cell and to cell population signaling with is limited by the number of proteins that the microfluidic chip the and of cell that have the allowed to of while to for the heterogeneity in phosphorylation in with in for signaling in different not with We developed regression to of the and to predict age of ex vivo cells. have been to signaling and of the of and by and to L. Quantitative combinations of predict model of signaling a for analysis of by regression of dynamic multivariate of this of that information content is within phosphorylation events in T cells L. Quantitative combinations of predict is to multivariate analysis and in and to and surface From the CD28 and CD27 expression markers in cellular age and to the model not the surface markers and and age in culture, to a of replicative senescence and of CD8+ T because of the culture to been that the and stimulation cells is to CD28 for cytotoxic for adoptive is a marker with the T-cell expression of been on T cells. the of stimulation, in the of cells capable of that of replicative senescence and of CD8+ T have loss of CD28 expression of CD28 expression on T cells is by receptor and and to cell before this model is limited in to the culture a that for expansion model the of early signaling following to and predict cellular age. have in early signaling events in in T from and in T from Jr., in the of proteins to in T the ex vivo culture expansion and major and from the model in accurately cellular age. is the to following to phosphorylation of the and phosphorylation of and T cell in T from lymphocyte signaling and and in in and and Jr., to are for the of proteins. have with age for the in in expression in the that this is and depends on the in We that the heterogeneity in expression and cellular is with cellular age. age and of phosphorylation and of a multivariate analysis this a high in to with and in T cell from heterogeneity in The multivariate model is to the age and the of a cell population for cells stimulation with multivariate signaling information with cellular model of signaling a for L. Quantitative combinations of predict of a on cell the of to the of the this is not a cell that we and the biomarkers are not in a to the age of the cells. of a a to the the model on the most informative markers of the for and predicts cellular age from We this model to a to age of the of a microfluidic chip the analysis of T cell by in the days and We of and a chip to the early phosphorylation events with in to informative in the regression prediction of population and days in to a cell that cells for of population The fixation and facilitate the of single-cell information with We the at the for in of the from with with