To identify small molecules which can block TNFR2-mediated Treg proliferation, we optimized the screening system based on our previous in vitro experiment setting of TNF-stimulated Treg proliferation (Chen, et al., 2007; Nie, et al., 2018). Splenocytes and LN cells from C57BL/6-Tg(Foxp3-DTR/EGFP)23.2Spar/Mmjax mice which allow unambiguously identification of Foxp3+ Treg populations based on EGFP expression (Lahl, et al., 2007) were used in the initial screening. As can be seen in Supplementary Fig. S1, the addition of TNF (5 ng/ml) to mixed lymphocytes resulted in raised EGFP-Foxp3 expression in CD4+ T cells, increased divided number of EGFP-Foxp3+ cells (monitored by the dilution of CellTrace™ Violet) and high expression of TNFR2, which could be markedly inhibited by the positive control etanercept (ETA), a combination of the extracellular domain of TNFR2 linked to the Fc portion of a human IgG1.
We then screened natural and synthetic small-molecule library in our lab (Kunming Institute of Botany, Chinese Academy of Sciences), which led to the identification of 1-phenyl-1,2,3,4-tetrahydro-isoquinoline (Alk2) possessing inhibitory activity on TNF-induced increase of EGFP-Foxp3 expressing cells in CD4 T cells (Supplementary Fig. S2). We further confirmed the activity of Alk2 on Tregs from wild type C57BL/6 mice. As shown in Fig. 1, under non-toxic concentrations (Supplementary Fig. S3), Alk2 significantly inhibited TNF-induced proliferation of Tregs in a concentration-dependent manner (p < 0.05 − 0.01), as assessed by the proportion of Foxp3+ cells in CD4+ T cells (Fig. 1A & B) and by the replication of CD4+Foxp3+ T cells (Fig. 1C & 1D). Furthermore, Alk2 significantly inhibited the elevation of TNFR2 expression on Tregs induced by TNF (Fig. 1E & F). We further confirmed the results with purified Tregs. As shown in Fig. 2A&B, Alk2 at 0.5 µM significantly inhibited TNF-induced proliferation of FACS-sorted Tregs (p < 0.01).
Previously, we found that TNF promoted the expansion and function of Tregs through its receptor TNFR2 (Chen, et al., 2007). To further examine if Alk2 inhibited Treg proliferation depended on TNFR2-pathway, high concentration (10 ng/ml) of IL-2 only was used to induce the expansion of Tregs. As shown in Fig. 2C&D, in contrast to the obvious inhibition of Alk2 (0.5 µM) on TNF-induced proliferation (Fig. 2A & 2B), Alk2 at 0.5 µM and even 1 µM did not show any inhibition on IL-2-induce proliferation of purified Tregs. These effects were further confirmed with TNFR2-deficient Tregs. As shown in Fig. 2E&F, high concentration of IL-2 was able to stimulate the proliferation of TNFR2-deficient Tregs which could be markedly inhibited by vindesine (Vin), an inhibitor of mitosis. However, Alk2 at 0.5 µM did not show any inhibition, albeit at 1 µM showed marginal inhibition. In another experiment (Supplementary Fig. S4), we assayed the expression of 4-1BB and CD25 on IL-2-stimulated TNFR2-deficient Tregs. In consistent with aforementioned results, Alk2 did not inhibit IL-2-induced elevation of 4-1BB and CD25 expression on Tregs. These results clearly indicated that Alk2 selectively inhibited TNFR2-mediated proliferation of Tregs.
To investigate if Alk2 interferes with TNF-TNFR2 pathway by directly blocking the binding of TNF to extracellular domain of TNFR2, we determined the effect of Alk2 on TNF-induced cell death of WEHI-13 VAR, since we found previously that the cytotoxicity of TNF on WEHI-13 VAR was mediated by TNFR2. The result was shown in Supplementary Fig. S5. TNF significantly decreased the cell viability by 60% as compared with the vehicle control group. ETA, but not Alk2 markedly inhibited the cytotoxic effect of TNF. This result suggested that Alk2 did not directly block extracellular TNF-TNFR2 binding. These data favored that idea that the inhibition of Alk2 on Treg proliferation depended on the blockade of intracellular TNFR2 signaling pathway.
In TME, increase of the ratio of effector T cells (Teffs) to Tregs is vital in the immune surveillance of malignancy and for a better prognosis of patients (Verma, et al., 2019). To examine the effect of Alk2 on the proliferation of Teffs, CD4 and CD8 T cells were stimulated with anti-CD3 antibody and anti-CD28 antibody. Result in Supplementary Fig S6 showed that Alk2 at 5 µM did not inhibit the proliferative responses of both CD4+ and CD8+ T cells in response to the stimulation with anti-CD3/28 antibodies. Furthermore, anti-CD3/28-induced Th1/Th2/Th17 cytokines secretion from T cells was not inhibited by Alk2 at 5 µM. Moreover, Alk2 (5 µM) did not inhibit the differentiation of TGF-β-induced Treg differentiation (Supplementary Fig. S7). It is worth noting that the IC50 of Alk2 in the inhibition of C8166 cell proliferation was 30.9 µM (Cheng, et al., 2008). Therefore, inhibition of Alk2 at 5 µM on TNF-induced expansion of Tregs was not based on the general cytotoxic effect of this compound. Instead, our data favored the idea that the effect of this compound is likely mediated by inhibition of TNF-TNFR2 signaling. The exact mechanism should be further clarified in the future study.
Alk2 belongs to the isoquinoline class which has a broad variety of biological activities (Shang, et al., 2020). To investigate if the isoquinoline itself has the Alk2-like activity, five isoquinoline analogs were tested. As shown in Supplementary Fig. S8, these compounds at 5 or 10 µM did not inhibit the replication of CD4+ Foxp3+ T cells and the up-regulation of TNFR2 expression on TNF-stimulated Tregs. Therefore, the isoquinoline itself has no activity on TNF-induced Treg proliferation. The above facts suggested that the 1-aryl substitution was required for the activity of Alk2.
To further study the structure-activity relationships (SARs), the effect of Alk2 analogs synthesized as reported previously (Cheng, et al., 2008) on TNF-induced proliferation of Tregs were examined. These compounds were divided into two types, namely, 6,7-dimethoxylated and 6,7-dihydroxylated derivatives. To this end, lymphocytes from C57BL/6-Tg(Foxp3-DTR/EGFP)23.2Spar/Mmjax mice were stained with CellTrace™ Violet, and then the cells were cultured with TNF, in the presence or absence of Alk2 analogs. The results, shown in Table 1, indicated that all 6,7-dimethoxylated derivatives showed similar or higher activity than the 6,7-dihydroxylated derivatives except for the pyridin-2-yl and 3-bromophenyl derivatives. The following SARs were discussed according to the above two skeletons.
Table 1
Inhibitory activities of 1-aryl-1,2,3,4-tetrahydro-isoquinolines on TNF-induced proliferation of Tregs.
No.
|
R
|
Inhibitory rate (%)
|
No.
|
R
|
Inhibitory rate (%)
|
1
|
|
36.3
|
23
|
|
27.9
|
2
|
|
25.4
|
24
|
|
30.8
|
3
|
|
25.1
|
25
|
|
22.9
|
4
|
|
24.3
|
26
|
|
25.7
|
5
|
|
24.3
|
27
|
|
22.9
|
6
|
|
25.6
|
28
|
|
24.7
|
7
|
|
32.6
|
29
|
|
25.4
|
8
|
|
20.6
|
30
|
|
28.3
|
9
|
|
40.7
|
31
|
|
24.7
|
10
|
|
26
|
32
|
|
17.2
|
11
|
|
33.9
|
33
|
|
31.1
|
12
|
|
39.1
|
34
|
|
33.0
|
13
|
|
29.9
|
35
|
|
28.1
|
14
|
|
19.4
|
36
|
|
9.9
|
15
|
|
26.2
|
37
|
|
25
|
16
|
|
28.4
|
38
|
|
24.3
|
17
|
|
24.9
|
39
|
|
17.6
|
18
|
|
27.5
|
‒
|
‒
|
‒
|
19
|
|
39.8
|
40
|
|
30.1
|
20
|
|
41.2
|
41
|
|
25.6
|
21
|
|
49.6
|
‒
|
‒
|
‒
|
22
|
|
42.3
|
‒
|
‒
|
‒
|
For the 6,7-dimethoxylated derivatives: the introduction of hetero atoms in aryl groups decreased the activity, e.g. the furan-2-yl (25.1%) and pyridin-2-yl (25.4) derivatives showed weaker activity than the phenyl derivative (36.3%). The position of halogens on the aryl group also influenced the activity. Different halogens (F, Cl, Br) with the ortho-substitution showed similar activity; whereas the chlorinated derivative showed higher activity than the bromated derivative when located at meta-position. Interestingly, the para-bromated derivative showed the highest activity with an inhibition ratio of 40.7%. Similarly, the para-methylated derivative (33.9%) also displayed higher activity than ortho-substitution (26.0%). In accordance with the previous report, the trifluoromethyl derivative showed increased effect than the methylated analogue, which might be due to the electron-withdrawing effects of the fluorine atoms. The introduction of methoxy groups on phenyl always decreased the activity, especially the para-methoxylated derivative showed the weakest activity (19.4%), markedly lower than the 2-methoxy and 2,4-dimethoxy derivatives. Compared with the non-substituted phenyl group, the introduction of nitro and cyano groups showed decreased activity. When the phenyl group was changed to be a naphthyl group, the inhibition was increased, suggesting that the large conjugated system was associated with the activity. For the naphthalen-1-yl derivatives, the 4-methoxylation showed the highest activity with inhibition ratio of 49.6%. The 6,7-dihydroxylated derivatives always showed similar or weaker activity than their respective 6,7-dimethoxylated derivative, except for the pyridin-2-yl and the meta-bromophenyl derivatives.
In conclusion, this study indicated that 1-aryl-1,2,3,4-tetrahydro-isoquinolines could inhibit TNF-induced Treg proliferation through interruption of intracellular TNFR2 signaling pathway. Therefore, these compounds may be useful as adjuvants for cancer immunotherapy and thus merit further research.