Tyrosine phosphatases regulate resistance to ALK inhibitors in ALK+ anaplastic large cell lymphoma

Abstract

Anaplastic large cell lymphomas (ALCLs) frequently carry oncogenic fusions involving the anaplastic lymphoma kinase (ALK) gene. Targeting ALK using tyrosine kinase inhibitors (TKIs) is a therapeutic option in cases relapsed after chemotherapy, but TKI resistance may develop. By applying genomic loss-of-function screens, we identified PTPN1 and PTPN2 phosphatases as consistent top hits driving resistance to ALK TKIs in ALK+ ALCL. Loss of either PTPN1 or PTPN2 induced resistance to ALK TKIs in vitro and in vivo. Mechanistically, we demonstrated that PTPN1 and PTPN2 are phosphatases that bind to and regulate ALK phosphorylation and activity. In turn, oncogenic ALK and STAT3 repress PTPN1 transcription. We found that PTPN1 is also a phosphatase for SHP2, a key mediator of oncogenic ALK signaling. Downstream signaling analysis showed that deletion of PTPN1 or PTPN2 induces resistance to crizotinib by hyperactivating SHP2, the MAPK, and JAK/STAT pathways. RNA sequencing of patient samples that developed resistance to ALK TKIs showed downregulation of PTPN1 and PTPN2 associated with upregulation of SHP2 expression. Combination of crizotinib with a SHP2 inhibitor synergistically inhibited the growth of wild-type or PTPN1/PTPN2 knock-out ALCL, where it reverted TKI resistance. Thus, we identified PTPN1 and PTPN2 as ALK phosphatases that control sensitivity to ALK TKIs in ALCL and demonstrated that a combined blockade of SHP2 potentiates the efficacy of ALK inhibition in TKI-sensitive and -resistant ALK+ ALCL.

Graphical abstract

Figure 1.

Genome-scale knockout screening reveals genes that meditate crizotinib resistance in anaplastic large cell lymphoma (ALCL). (A) Experimental design of the crizotinib resistance screen in ALK⁺ ALCL cell lines using the GecKO v2 library. (B) Genes ranked by MAGeCK Robust Rank Aggregation enrichment score in crizotinib over DMSO-treated cells, highlighting the top 10 of candidate genes acquired following specific filtering criteria based on the ranking list (see "Materials and methods"). ***FDR ≤0.01, **FDR ≤0.1, *FDR ≤ 0.2. (C) Deletion of PTPN1 or PTPN2 in ALK⁺ ALCL cell lines. TS and SU-DHL-1 cells were transduced with 2 different sgRNAs targeting either PTPN1 or PTPN2. Cells were treated with 80 nM or 300 nM crizotinib to evaluate changes in PTPN1 or PTPN2 expression. Western blotting was performed on cell lysates probed with the indicated antibodies. β-actin or β-tubulin were used as loading controls. One representative experiment of 4 is shown. (D) Cell viability assay was performed on TS and SU-DHL-1 WT or transduced cells with either PTPN1 or PTPN2 targeting sgRNAs undergoing crizotinib treatment. Data are represented as mean ± standard deviation (SD) of technical triplicates; **P < .01, ***P < .001, ****P < .0001, 2-way ANOVA followed by Dunnett's multiple comparisons test. (E) Growth curves of WT, PTPN1 KO, or PTPN2 KO ALK⁺ cells (TS and SU-DHL-1) treated with 80 nM crizotinib. Data are represented as means ± SD of technical triplicates; ****P < .0001, 2-way ANOVA followed by Tukey's multiple comparisons test. (F) PTPN1 KO or PTPN2 KO TS cells were subcutaneously injected into NSG mice. The mice were treated with crizotinib (75 mg/kg daily) by oral gavage. Data are represented as means ± SD of 4-6 mice in the control groups and 8 mice in the crizotinib-treated groups; ****P < .0001, 2-way ANOVA followed by Tukey's multiple comparisons test.

Figure 1.

Genome-scale knockout screening reveals genes that meditate crizotinib resistance in anaplastic large cell lymphoma (ALCL). (A) Experimental design of the crizotinib resistance screen in ALK⁺ ALCL cell lines using the GecKO v2 library. (B) Genes ranked by MAGeCK Robust Rank Aggregation enrichment score in crizotinib over DMSO-treated cells, highlighting the top 10 of candidate genes acquired following specific filtering criteria based on the ranking list (see "Materials and methods"). ***FDR ≤0.01, **FDR ≤0.1, *FDR ≤ 0.2. (C) Deletion of PTPN1 or PTPN2 in ALK⁺ ALCL cell lines. TS and SU-DHL-1 cells were transduced with 2 different sgRNAs targeting either PTPN1 or PTPN2. Cells were treated with 80 nM or 300 nM crizotinib to evaluate changes in PTPN1 or PTPN2 expression. Western blotting was performed on cell lysates probed with the indicated antibodies. β-actin or β-tubulin were used as loading controls. One representative experiment of 4 is shown. (D) Cell viability assay was performed on TS and SU-DHL-1 WT or transduced cells with either PTPN1 or PTPN2 targeting sgRNAs undergoing crizotinib treatment. Data are represented as mean ± standard deviation (SD) of technical triplicates; **P < .01, ***P < .001, ****P < .0001, 2-way ANOVA followed by Dunnett's multiple comparisons test. (E) Growth curves of WT, PTPN1 KO, or PTPN2 KO ALK⁺ cells (TS and SU-DHL-1) treated with 80 nM crizotinib. Data are represented as means ± SD of technical triplicates; ****P < .0001, 2-way ANOVA followed by Tukey's multiple comparisons test. (F) PTPN1 KO or PTPN2 KO TS cells were subcutaneously injected into NSG mice. The mice were treated with crizotinib (75 mg/kg daily) by oral gavage. Data are represented as means ± SD of 4-6 mice in the control groups and 8 mice in the crizotinib-treated groups; ****P < .0001, 2-way ANOVA followed by Tukey's multiple comparisons test.

Figure 2.

PTPN1 KO and PTPN2 KO cells show increased phosphorylation of SHP2, STAT3 and ERK compared with WT cells following crizotinib treatment. (A-B) Western blot analysis of PTPN1 KO or PTPN2 KO TS cells treated with the indicated concentrations of crizotinib for 4 hours. Cell lysates were blotted with the indicated antibodies. β-actin was used as a loading control. One representative experiment of 3 with comparable results is shown. (C) Western blot analysis performed on JB6 cells transduced with a doxycycline-inducible lentivirus expressing PTPN1 or PTPN2. Cells were induced with doxycycline for 24 or 48 hours and collected. Cell lysates were blotted with indicated antibodies. β-actin was used as a loading control. (D) Quantification of PTPN1, PTPN2, pALK (Tyr1604), and pALK (Tyr1278) expression levels following PTPN1 or PTPN2 overexpression induced by doxycycline in JB6 as in panel C. Data are represented as means ± SD of average of the 24- and 48-hour time points. The P value was determined by multiple t tests.

Figure 3.

NPM-ALK interacts with PTPN1 and PTPN2. (A) Co-IP assay was performed on TS cells treated with 300 nM crizotinib for 3 hours. Co-IP was performed with anti-ALK antibody (IP ALK) or the corresponding isotype control (IgG) and analyzed by western blot probed with the indicated antibodies. One representative experiment out of 2 is shown. (B) Co-IP was performed on 293T cells transfected with NPM-ALK or the mock vector (-) and collected after 48 hours. Co-IP was performed with anti-ALK antibody (IP ALK) or the corresponding isotype control (IgG) and analyzed by western blot probed with the indicated antibodies. One representative experiment out of 2 is shown. (C-D) Representative images of in situ PLA (red) between NPM-ALK and PTPN1 (C) and NPM-ALK and PTPN2 (D) in TS and JB6 cell lines. WT TS and JB6 cells treated for 12 hours with TL13-112 100 nM were used as control to specifically reduced the anti-ALK antibody binding; PTPN1 KO and PTPN2 KO JB6 and TS cells were used as negative controls to remove anti-PTPN1 and PTPN2 antibody binding. (E) Quantification of PLA between ALK and PTPN1 and ALK and PTPN2 (C). Data are represented as mean ± SD; *P < .05, Student t test. TCL, total cell lysates.

Figure 4.

SHP2 phosphorylation is increased in ALK⁺ ALCL. (A) Western blot analysis performed on human ALK⁺ ALCL (TS, SU-DHL-1, JB6, and COST), ALK- ALCL (OCI-Ly13.2, DL-40, and MAC2A) and healthy donor T cells and PBMCs with the indicated antibodies. (B) Western blot analysis performed on 293T cells transduced with a doxycycline-inducible lentivirus expressing PTPN1 or PTPN2. 293T-HEK cells were transfected with NPM-ALK and/or SHP2 vectors as indicated, induced with doxycycline for 48 hours, and then collected. Cell lysates were blotted with indicated antibodies. β-Actin was used as a loading control. (C) Competitive cell growth assay performed on TS and JB6 cells transduced with a retrovirus expressing empty vector, SHP2^wt, and SHP2^E76K or SHP2^D61Y mutant vectors cultured in presence of 40 nM crizotinib for indicated time. Data are represented as means ± SD of technical triplicates; **P = .0061 ****P < .0001, 2-way ANOVA followed by Dunnett's multiple comparisons test.

Figure 5.

SHP2 inhibition overcomes crizotinib resistance in ALK⁺ ALCL. (A-B) Western blot analysis performed on WT, PTPN1 KO, and PTPN2 KO JB6 cells treated with 80 nM of crizotinib and 1 or 5 μM of SHP099 in single or in combination for 1 hour. Cell lysates were blotted with indicated antibodies. One representative experiment out of 3 is shown. (C) Dose response curves of WT, PTPN1, and PTPN2 KO TS and JB6 cells incubated in increasing concentration of crizotinib in single or combination with SHP099 inhibitor at the indicated concentration for 72 hours. Data are represented as mean ± SD of technical triplicates; **P < .01, ****P < .0001, 2-way ANOVA followed by Dunnett's multiple comparisons test. (D-E) Combenefit synergy plot and Bliss Synergy Score of combined crizotinib and SHP099 in WT, PTPN1 KO, and PTPN2 KO TS (D), or JB6 (E) cell lines. Blue: synergy, green: additivity, and red: antagonism.

Figure 5.

SHP2 inhibition overcomes crizotinib resistance in ALK⁺ ALCL. (A-B) Western blot analysis performed on WT, PTPN1 KO, and PTPN2 KO JB6 cells treated with 80 nM of crizotinib and 1 or 5 μM of SHP099 in single or in combination for 1 hour. Cell lysates were blotted with indicated antibodies. One representative experiment out of 3 is shown. (C) Dose response curves of WT, PTPN1, and PTPN2 KO TS and JB6 cells incubated in increasing concentration of crizotinib in single or combination with SHP099 inhibitor at the indicated concentration for 72 hours. Data are represented as mean ± SD of technical triplicates; **P < .01, ****P < .0001, 2-way ANOVA followed by Dunnett's multiple comparisons test. (D-E) Combenefit synergy plot and Bliss Synergy Score of combined crizotinib and SHP099 in WT, PTPN1 KO, and PTPN2 KO TS (D), or JB6 (E) cell lines. Blue: synergy, green: additivity, and red: antagonism.

Figure 6.

A combination of crizotinib and SHP099 overcomes TKI resistance in ALK⁺ ALCL in vivo. (A) Growth curves of tumor xenografts of PTPN1 KO JB6 cells injected subcutaneously in NSG mice treated with crizotinib (75 mg/kg daily), SHP099 (75 mg/kg daily), or a combination of both drugs. Data are represented as means ± SD; *P < .05, ***P < .001, ****P < .0001, 2-way ANOVA followed by Tukey's multiple comparisons test; n = 3 mice for control group, n = 10 mice for crizotinib treated group, n = 8 mice for SHP099-treated group, and n = 9 mice for combo group. (B) Representative immunohistochemistry for Ki-67 and cleaved caspase-3 performed on crizotinib-treated or crizotinib + SHP099–treated PTPN1 KO tumors. Scale bar, 50 μm. (C) Quantification of Ki-67 or cleaved caspase-3⁺ cells in PTPN1 KO tumors treated as in Figure 6A. Significance was determined by unpaired, 2-tailed Student t test, **P < .01, ****P < .0001. (D) Western blot analysis performed on lysates from PTPN1 KO tumors obtained from mice treated with the indicated drugs shows marked downregulation of ERK phosphorylation in the crizotinib plus SHP099 treatment. Two representative tumors for each treatment are shown.

Figure 7.

Patient samples resistant to ALK TKI show downregulation of PTPN1 and PTPN2 and upregulation of SHP2. (A) Representative hematoxylin-eosin and immunohistochemistry stainings of ALK, CD30, PTPN1, and PTPN2 in 1 ALK⁺ case expressing low PTPN1 and low PTPN2 (top) and 1 ALK⁺ ALCL case expressing high PTPN1 and high PTPN2 (bottom). Tumor cells are identified by ALK and CD30 stainings. High expression was determined by an H score >200. Scale bar, 100 μm. (B) Percentage of cases with high PTPN1 or PTPN2 protein expression of PTPN1 and PTPN2 evaluated by immunohistochemistry in human T-cell lymphoma subtypes: ALK⁺ ALCL (n = 30), ALK⁻ ALCL (n = 18), AITL (n = 10), and peripheral T-cell lymphoma (n = 10). High expression was determined by an H score >200. (C) RNA-seq data from chemotherapy-relapsed/refractory (n = 2) and ALK TKI–resistant (n = 2) patients to evaluate mRNA expression of PTPN1, PTPN2, and PTPN11 (SHP2). ALK mutation status was analyzed by sequencing following a biopsy of relapsed lymphoma samples. The 2 patients that relapsed after chemotherapy showed WT ALK. Of the 2 patients relapsed during ALK TKI, 1 patient was WT, and 1 patient carried the L1196M mutation in ALK. See supplemental Figure 12 for treatment details. CPM, counts per million.