Persistence of tumor infiltrating lymphocytes in adoptive immunotherapy correlates with telomere length

Abstract

Transfer of autologous tumor-specific tumor infiltrating lymphocytes (TILs) in adoptive immunotherapy can mediate the regression of tumor in patients with metastatic melanoma. In this procedure, TILs from resected tumors are expanded in vitro, then administered to patients and further stimulated to proliferate in vivo by the administration of high dose IL-2. After in vitro expansion, TILs are often dominated by a few specific clonotypes, and recently it was reported that the persistence in vivo of one or more of these clonotypes correlated with positive therapeutic response. We and others have previously shown that repeated in vitro stimulation and clonal expansion of normal human T lymphocytes results in progressive decrease in telomerase activity and shortening of telomeres, ultimately resulting in replicative senescence. In the studies reported here, we therefore compared telomerase activity and telomere length in persistent and nonpersistent TIL clonotypes before transfer in vivo, and found a correlation between telomere length and clonal persistence. We also observed that TILs proliferate extensively in vivo in the days after transfer, but fail to induce substantial telomerase activity, and undergo rapid decreases in telomere length within days after transfer. Thus, in vivo loss of telomeres by clonotypes that have the shortest telomeres at the time of administration may drive these clones to replicative senescence, whereas cells with longer telomeres are able to persist and mediate antitumor effects. These findings are relevant both to predicting effectiveness of adoptive immunotherapy and in deriving strategies for improving effectiveness by sustaining telomere length.

FIGURE 1

Telomere length of persistent and nonpersistent clonotypes. Each dot represents the average (when more than 1 was assayed) of persistent or nonpersistent clonotypes from 1 patient. A, Telomere length of persistent and nonpersistent clonotypes across individual patients. Mean length of persistent clonotypes was 6.6 kb and nonpersistent was 5.0 kb (P<0.04). For this calculation, because some patients had more than 1 persistent or nonpersistent clonotype, the average telomere length of persisters or nonpersisters in that patient was calculated and used in statistical analysis. When telomere lengths for each individual clonotype were used for analysis, mean length of persistent clonotypes was 6.4 kb and mean of nonpersistent clonotypes was 5.0 kb (P<0.02). A total of 27 clonotypes from 14 patients were measured. B, Pair-wise comparison of persistent and nonpersistent clonotypes from the same patient (P<0.02). Mean lengths of persistent or nonpersistent clonotypes for an individual patient were used in this pair-wise comparison. This represents 14 clonotypes from 5 patients.

FIGURE 2

Telomere length of persistent clonotype after transfer in vivo. Telomere lengths of the persistent clonotype in 4 patients were measured by 2 color flow-FISH. Each showed an initial decline in length, followed by maintenance.

FIGURE 3

Ki-67 expression of TILs after transfer in vivo. This representative experiment presents the percentages of cells with a persistent TIL clonotype that express Ki-67 increase in the few days after transfer, indicating an early and transient peak of proliferative activity.

FIGURE 4

Telomerase activity in TILs. TILs after expansion had very low telomerase activity, in vitro and in vivo, and did not up-regulate telomerase upon restimulation. TILs or freshly isolated peripheral blood leukocytes were stimulated with anti-CD3 and APC as described in Materials and Methods. CD8-enriched populations of peripheral blood leukocyte were isolated after stimulation; and clonotypic TIL populations were isolated by tetramer binding or TCR Vβ expression. Lysates from these populations were assayed for telomerase activity as described, and activity is expressed in relative units on the basis of a standard included in each assay. Telomerase posttransfer in TIL2 and TIL6 were not examined due to unavailability of samples.