Characterization of Matrigel, fibrin, and fibrin/Matrigel

To evaluate whether the fibrin/Matrigel is suitable for 3D invasion assays, we first cultured MRC-5 fibroblasts in Matrigel, fibrin, and fibrin/Matrigel. As shown in Fig. 1(a), the fibroblasts in 3 mg/ml Matrigel became rounded with F-actin surrounding the nucleus. Fibroblasts would not migrate with such a morphology. In contrast, in 5 mg/mL fibrin and the fibrin/Matrigel, fibroblasts extended into a spindle shape with elongated F-actin. We speculated that the stiffness of 3 mg/ml Matrigel might be too low for the attachment and extension of fibroblasts.2828. D. E. Discher, P. Janmey, and Y.-L. Wang, Science 310, 1139–1143 (2005). https://doi.org/10.1126/science.1116995 Moreover, the gene profiles of cells cultured in fibrin or fibrin/Matrigel might be different from those of cells in Matrigel.2929. W. Wahbi, E. Naakka, K. Tuomainen, I. Suleymanova, A. Arpalahti, I. Miinalainen, J. Vaananen, R. Grenman, O. Monni, A. Al-Samadi, and T. Salo, Exp. Cell Res. 389, 111885 (2020). https://doi.org/10.1016/j.yexcr.2020.111885 Because pure fibrin might not resemble the ECM constituents in the TME, we decided to use fibrin/Matrigel in the following experiments.We used FIB-SEM to observe the microstructures of these hydrogels. The micrographs in Fig. 1(b) show that Matrigel exhibited obvious sheet structures, whereas fibrin formed mostly fibrillar structures. We measured the average pore sizes of Matrigel, fibrin, and fibrin/Matrigel in the SEM images to be 2.06 ± 1.36, 0.79 ± 0.29, and 1.34 ± 0.87 μm2, respectively. The average pore densities of Matrigel, fibrin, and fibrin/Matrigel were 0.09 ± 0.04, 0.56 ± 0.17, and 0.23 ± 0.08 μm−2, respectively. Compared with Matrigel, the addition of fibrin slightly reduced the pore size and increased the pore density. We found that in fibrin/Matrigel, the fibrin fiber was adhered to the surface of the Matrigel sheet structures. The fibrillar structures of fibrin and fibrin/Matrigel could also be necessary for 3D cancer cell invasion, as proposed by Berger et al.3030. A. J. Berger, K. M. Linsmeier, P. K. Kreeger, and K. S. Masters, Biomaterials 141, 125–135 (2017). https://doi.org/10.1016/j.biomaterials.2017.06.039AFM analysis showed that the mean stiffness values of 3 mg/ml Matrigel and 5 mg/ml fibrin were 6 ± 2 and 258 ± 25 Pa, respectively, whereas that of fibrin/Matrigel was 105 ± 44 Pa. The stiffness distributions in each type of hydrogel are shown in supplementary material Fig. S1. The AFM analysis suggested that the fibrillar structures of fibrin/Matrigel shown in Fig. 1(b) provided additional physical support to the hydrogel.We also performed the alamarBlue assay to quantify the viabilities of A549 cancer cells and MRC-5 fibroblasts in the three types of hydrogels. The data in supplementary material Fig. S2(a) show that both types of cells preserved high viabilities in these hydrogels. The invasiveness of cancer cells and fibroblasts quantified by the transwell assay is shown in supplementary material Fig. S2(b). The invasion capabilities of monocultured cancer cells or fibroblasts were similar in the three types of hydrogels.Collagen is widely used in various hydrogels for 3D cell or tissue culture.1111. Y. Li and E. Kumacheva, Sci. Adv. 4(4), eaas8998 (2018). https://doi.org/10.1126/sciadv.aas8998 Nonetheless, if fibroblasts exist in the culture system, the amounts of collagen produced by the fibroblasts could be a major variable reflecting the activity of fibroblasts; therefore, the exogenous collagen may be considered an undesirable background factor. In addition, myofibroblasts could exert deformation forces along fibrillar collagen ECM up to hundreds of micrometers.3131. P. Pakshir, M. Alizadehgiashi, B. Wong, N. M. Coelho, X. Chen, Z. Gong, V. B. Shenoy, C. A. McCulloch, and B. Hinz, Nat. Commun. 10, 1850 (2019). https://doi.org/10.1038/s41467-019-09709-6 ECM deformation could affect the migration and invasion of nearby cells. Considering these points, we suggest that the fibrin/Matrigel used in the present study should be more suitable for assessing 3D cancer cell invasion under the effect of fibroblasts. Moreover, although the properties of Matrigel could vary from batch to batch, at least stiffness could be controlled by adjusting the amount of fibrin in fibrin/Matrigel.

Cancer cell spheroids in fibrin/Matrigel without and with fibroblasts

In the co-culture experiments, we cultured one cancer cell spheroid with fibroblasts in a well containing fibrin/Matrigel and then observed the cell spheroids at 0 h and 72 h via confocal microscopy [Fig. 2(a)]. Figure 2(b) shows the intensity projection fluorescence and bright-field images of cancer cell spheroids in fibrin/Matrigel with and without the co-culture of MRC-5 fibroblasts. In these images, only the cancer cells were loaded with the fluorescent dye. Compared with our previous results on cancer cell spheroids and fibroblasts in Matrigel,1616. H.-H. Hou, H.-J. Pan, W.-Y. Liao, C.-H. Lee, and C.-J. Yu, Int. J. Cancer 147(9), 2587–2596 (2020). https://doi.org/10.1002/ijc.33127 the images in Fig. 2(b) showed much more obvious cancer cell invasion with the co-culture of fibroblasts, demonstrating better potential of fibrin/Matrigel for spheroid invasion assays.In Fig. 2(c), the mean projection area of monoculture cancer cell spheroids at 72 h normalized to the initial value was 1.8 ± 0.2. We controlled the initial cancer cell numbers to be ∼1000 per well. The coefficient of variation of spheroid areas at 0 h for 327 monoculture spheroids was 18.4%. With the co-culture of fibroblasts (concentration: 1 × 106 ml−1) for 72 h, the normalized spheroid area was 4.6 ± 1.0, i.e., 2.5-fold of that in the monoculture. As an evaluation of the invasiveness of the cancer cells, we used the inverse of spheroid circularity (1/C value) to quantify the irregularity of spheroid circumferences. Higher cell invasion resulted in more protrusions of the spheroid, which increased the 1/C value. For monocultured spheroids, the average 1/C value after 72 h in fibrin/Matrigel was 3.2 ± 1.4. The co-culture with fibroblasts significantly increased the average 1/C value to 31.5 ± 19.5, indicating more cancer cell invasion in the co-culture [Fig. 2(d)]. This result agrees with the well-known fibroblast-assisted cancer cell invasion.4,5,7,32,334. J. Mahale, G. Smagurauskaite, K. Brown, A. Thomas, and L. M. Howells, Int. J. Cancer 138(1), 30–44 (2016). https://doi.org/10.1002/ijc.294475. N. K. Altorki, G. J. Markowitz, D. Gao, J. L. Port, A. Saxena, B. Stiles, T. McGraw, and V. Mittal, Nat. Rev. Cancer 19, 9–31 (2019). https://doi.org/10.1038/s41568-018-0081-97. K. Y. Wong, A. H.-K. Cheung, B. Chen, W. N. Chan, J. Yu, K. W. Lo, W. Kang, and K. F. To, Int. J. Cancer 151, 1195–1215 (2022). https://doi.org/10.1002/ijc.3412732. A. Abulaiti, Y. Shintani, S. Funaki, T. Nakagiri, M. Inoue, N. Sawabata, M. Minami, and M. Okumura, Lung Cancer 82, 204–213 (2013). https://doi.org/10.1016/j.lungcan.2013.08.00833. S. Chen, A. Giannakou, S. Wyman, J. Gruzas, J. Golas, W. Zhong, C. Loreth, L. Sridharan, T.-T. Yamin, M. Damelin, and K. G. Geles, Proc. Natl. Acad. Sci. U. S. A. 115, E11671–E11680 (2018). https://doi.org/10.1073/pnas.1803718115 The major pro-invasion factors produced by fibroblasts include the hepatocyte growth factor, interleukin-6, transforming growth factor-β, and matrix metallopeptidases (MMPs). On the other hand, fibroblasts or cancer cells might produce collagen and other proteins to modify the structure and stiffness of the ECM, which could affect the cancer cell invasion and progression with complicated mechanisms.34,3534. C. Bonnans, J. Chou, and Z. Werb, Nat. Rev. Mol. Cell Biol. 15, 786–801 (2014). https://doi.org/10.1038/nrm390435. C. T. Mierke, Rep. Prog. Phys. 82(6), 064602 (2019). https://doi.org/10.1088/1361-6633/ab1628 Because higher stiffness could accompany a denser fibrillar network, higher matrix stiffness does not necessarily promote cell invasion.3636. A. Micalet, E. Moeendarbary, and U. Cheema, “3D in vitro models for investigating the role of stiffness in cancer invasion,” ACS Biomater. Sci. Eng. (published online) (2021). https://doi.org/10.1021/acsbiomaterials.0c01530 To reveal how the ECM's mechanical properties influence cancer cell invasion under the effects from co-cultured fibroblasts, more experimental work and numerical simulations are required.Another mechanism by which co-cultured fibroblasts improve cancer cell invasiveness could be fibroblast-led invasion.15,16,37–3915. A. Labernadie, T. Kato, A. Brugués, X. Serra-Picamal, S. Derzsi, E. Arwert, A. Weston, V. González-Tarragó, A. Elosegui-Artola, L. Albertazzi, J. Alcaraz, P. Roca-Cusachs, E. Sahai, and X. Trepat, Nat. Cell Biol. 19(3), 224–237 (2017). https://doi.org/10.1038/ncb347816. H.-H. Hou, H.-J. Pan, W.-Y. Liao, C.-H. Lee, and C.-J. Yu, Int. J. Cancer 147(9), 2587–2596 (2020). https://doi.org/10.1002/ijc.3312737. C. Gaggioli, S. Hooper, C. Hidalgo-Carcedo, R. Grosse, J. F. Marshall, K. Harrington, and E. Sahai, Nat. Cell Biol. 9, 1392–1400 (2007). https://doi.org/10.1038/ncb165838. S. Neri, G. Ishii, H. Hashimoto, T. Kuwata, K. Nagai, H. Date, and A. Ochiai, Int. J. Cancer 137, 784–796 (2015). https://doi.org/10.1002/ijc.2946439. Y. Attieh, A. G. Clark, C. Grass, S. Richon, M. Pocard, P. Mariani, N. Elkhatib, T. Betz, B. Gurchenkov, and D. M. Vignjevic, J. Cell Biol. 216(11), 3509–3520 (2017). https://doi.org/10.1083/jcb.201702033 Figure 2(e) (Multimedia view) shows time-lapse confocal microscopy images of the region near the rim of an A549 cell spheroid. Fibroblasts (yellow) led the invasion of A549 cells (red) into the fibrin/Matrigel. This result confirmed that fibroblast-led invasion also played a role in fibroblast-enhanced cancer cell invasion in our co-culture system.

Effects of anti-fibrotic drugs on cancer cell spheroids co-cultured with fibroblasts in fibrin/Matrigel

The results in Fig. 2 suggest that fibroblasts might enhance tumor growth and cancer cell invasion in fibrin/Matrigel. Therefore, to know how the anti-fibrotic drugs affected this cancer cell–fibroblast co-culture system, we used various concentrations of NTD and PFD to treat cancer cell spheroids co-cultured with fibroblasts (concentration: 1 × 106 ml−1) in fibrin/Matrigel and measured the area and 1/C values of the spheroids. The concentration ranges of NTD and PFD were used according to the report of Knüppel et al.4040. L. Knuppel, Y. Ishikawa, M. Aichler, K. Heinzelmann, R. Hatz, J. Behr, A. Walch, H. P. Bachinger, O. Eickelberg, and C. A. Staab-Weijnitz, Am. J. Respir. Cell Mol. Biol. 57(1), 77–90 (2017). https://doi.org/10.1165/rcmb.2016-0217OC The results in Fig. 3(a) show that the treatment of 0.5 or 1.0 μM NTD for 72 h significantly reduced both the area and 1/C value of cancer cell spheroids (P C value were still larger than those in the monoculture condition [Figs. 2(c) and 2(d)]. In contrast, 100, 500, and 1000 μM PFD showed no effect on the spheroid areas and 1/C values, as shown in Fig. 3(b).The cell viability (alamarBlue) assays confirmed that both 0.5 μM NTD and 1000 μM PFD did not reduce the viability of MRC-5 fibroblasts in 72 h (Fig. S3 in supplementary material). On the other hand, 50 or 100 μM CDDP significantly diminished the viability of both A549 cells and MRC-5 fibroblasts. Therefore, it would be interesting to know how the combination of NTD or PFD with CDDP affected spheroid growth and cancer cell invasion in the fibrin/Matrigel system.

Efficacy of CDDP plus anti-fibrotic drugs on cancer cell spheroids with and without fibroblast co-culture

We then evaluated the efficacy of CDDP in combination with NTD or PFD on A549 cell spheroids in fibrin/Matrigel with and without fibroblasts. The duration of anti-fibrotic drug pretreatment was 24 h, after which CDDP was applied for another 48 h [Fig. 4(a)]. All the results were normalized to those obtained right before the pretreatment (0 h) on the same spheroid. The data in Figs. 4(b)–4(d) show that without any treatment, the areas of monocultured A549 cell spheroid increased by nearly 90% after 72 h of culture. The treatment of 100 μM CDDP inhibited the growth of spheroids almost completely in the same period. The addition of 0.5 μM NTD did not show a measurable change in the results of CDDP treatment. The 1/C values of monocultured cancer cell spheroids were mostly between 2 and 4, and CDDP treatment reduced the 1/C values to ∼2. These results suggest that the invasion of A549 cancer cells was not obvious in fibrin/Matrigel without fibroblasts. The addition of NTD did not affect the 1/C values for 50-μM or 100-μM CDDP treatments. We repeated the above experiments using CDDP and PFD, and the data in Figs. S4(a)–S4(c) in the supplementary material show that the addition of 1000 μM PFD did not change the effects of CDDP on the growth and the invasiveness of monocultured A549 cell spheroids.Next, we evaluated the efficacy of CDDP in combination with NTD or PFD on A549 cancer cell spheroids co-cultured with fibroblasts. The concentration of fibroblasts in the hydrogel was 1 × 106 ml−1. Figure 4(e) shows that the addition of 0.5 μM NTD significantly reduced the fibroblast-assisted invasion of cancer cells in the co-culture system. According to the data in Figs. 4(f) and 4(g), the effects of 0.5 μM NTD without CDDP on the spheroid area and 1/C value were close to those of 100 μM CDDP without NTD in the co-culture system. The treatment of 0.5 μM NTD + 50 μM CDDP suppressed the spheroid area and 1/C value to those under the 100 μM CDDP treatment, suggesting that the combination of NTD might lead to a reduction in the CDDP dosage while preserving the antitumor effects. Moreover, the addition of 0.5 μM NTD to 50 μM or 100 μM CDDP reduced the spheroid areas and 1/C values more than CDDP alone.We also evaluated the effect of the combination of PFD and CDDP on A549 cell spheroids co-cultured with fibroblasts. Figures S4(d)–S4(f) in the supplementary material show that the addition of 1000 μM PFD to 50 μM or 100 μM CDDP did not exhibit observable effects on the spheroid area and 1/C values. Therefore, we hypothesized that the addition of 1000 μM PFD did not improve the effect of CDDP treatment on the co-culture system.Although there have been many studies using 2D culture to explore the combined effects of anti-fibrotic drugs with anticancer drugs on CAFs and cancer cells,41–4441. M. Mediavilla-Varela, K. Boateng, D. Noyes, and S. J. Antonia, BMC Cancer 16, 176 (2016). https://doi.org/10.1186/s12885-016-2162-z42. G. E. Shochet, L. Israeli-Shani, M. Koslow, and D. Shitrit, Lung Cancer 96, 7–14 (2016). https://doi.org/10.1016/j.lungcan.2016.03.01343. M. Gabasa, R. Ikemori, F. Hilberg, N. Reguart, and J. Alcaraz, Br. J. Cancer 117, 1128–1138 (2017). https://doi.org/10.1038/bjc.2017.27044. T. Yamanaka, N. Harimoto, T. Yokobori, R. Muranushi, K. Hoshino, K. Hagiwara, D. Gantumur, T. Handa, N. Ishii, M. Tsukagoshi, T. Igarashi, H. Tanaka, A. Watanabe, N. Kubo, K. Araki, and K. Shirabe, Br. J. Cancer 122, 986–994 (2020). https://doi.org/10.1038/s41416-020-0744-7 no work directly compared NTD and PFD on the proliferation and invasion of cancer cells, even in a 2D co-culture system. In the present work, for monocultured A549 cell spheroids in fibrin/Matrigel, the effects of CDDP and CDDP plus NTD or PFD on reducing the spheroid areas did not differ significantly. However, with the presence of fibroblasts, the combination of 50 μM CDDP and 0.5 μM NTD could achieve a size-reduction effect similar to that of 100 μM CDDP alongside an even better invasion-suppression effect than that of 100 μM CDDP. In contrast, the combination of CDDP and 1000 μM PFD did not show any observable variations in the effects of CDDP. This result is different from that obtained with the combination of 10 μM CDDP and 2.7 mM PFD in a conventional 2D co-culture system, wherein the combined treatment did increase the death ratio of cells treated with CDDP alone.4141. M. Mediavilla-Varela, K. Boateng, D. Noyes, and S. J. Antonia, BMC Cancer 16, 176 (2016). https://doi.org/10.1186/s12885-016-2162-z Another study using A549 cell spheroids co-cultured with fibroblasts in 3D collagen gels suggested that 2.7 mM PFD could reduce the expression of α-smooth muscle actin in fibroblasts.4545. H. Takahashi and Y. Kikuchi, Biomater. Sci. 9, 4448–4458 (2021). https://doi.org/10.1039/D1BM00629K However, in both studies, the effects of PFD on cancer cell invasiveness were not reported. Fujiwara et al. showed that in 2D culture PFD alone could suppress fibroblast activities and epithelial to mesenchymal transition of A549 cancer cells induced by the fibroblast conditioned medium.4646. A. Fujiwara, S. Funaki, E. Fukui, K. Kimura, T. Kanou, N. Ose, M. Minami, and Y. Shintani, Sci. Rep. 10, 10900 (2020). https://doi.org/10.1038/s41598-020-67904-8 However, because of the high concentration of PFD (2.7 mM), the invasion and growth of A549 cancer cells were inhibited even without the fibroblast conditioned medium. We hypothesized that in our experiments, the addition of 1000 μM PFD might have only a negligible effect on the interactions between cancer cells and fibroblasts, even under the treatment of 100 μM CDDP. Considering that the measured PFD concentrations in the blood of adult patients were less than 220 μM,4747. L. M. Barranco-Garduno, I. Buendía-Roldan, J. J. Rodriguez, R. Gonzalez-Ramírez, A. N. Cervantes-Nevarez, J. C. Neri-Salvador, M. d C. Carrasco-Portugal, G. Castaneda-Hernandez, K. Martinez-Espinosa, M. Selman, and F. J. Flores-Murrieta, Heliyon 6, e05279 (2020). https://doi.org/10.1016/j.heliyon.2020.e05279 our results could be closer to the conditions in living tissues.

Efficacy of anti-fibrotic drugs on fibroblasts in fibrin/Matrigel

The results in Fig. 4 indicate that for the monocultured A549 cell spheroids, the addition of anti-fibrotic drugs did not change the efficacy of CDDP on spheroid growth and cancer cell invasion. In the co-culture system, the anti-fibrotic drug, NTD, increased the efficacy of CDDP, but PFD did not show a similar effect. Because the co-culture of fibroblasts significantly enhanced spheroid growth and cancer cell invasion (Fig. 2), we assumed that NTD could suppress the activities of fibroblasts in the co-culture system more significantly than PFD. Moreover, the data in Fig. 4(g) suggest that 0.5 μM NTD could reduce the 1/C value more than 50 μM CDDP. Therefore, comparing the effects of NTD and PFD on fibroblasts in fibrin/Matrigel might be useful in revealing more details about the results shown in Fig. 4.We used confocal microscopy to observe the morphological variations of MRC-5 fibroblasts in fibrin/Matrigel. The medium was conditioned by A549 cells (2 × 105 ml−1, 24 h of culture). Figure 5(a) shows that the 0.5 μM NTD treatment for 48 h substantially changed the morphology of fibroblasts, and the actin filaments did not increase in length as those in fibroblasts treated with PFD. As further verification of the different effects between NTD and PFD on fibroblasts, we also treated normal human lung fibroblasts (NHLFs) in fibrin/Matrigel. Because of the slower response of NHLFs, the duration of treatment was extended to 72 h. The effects of NTD and PFD on NHLFs were similar to those on MRC-5 fibroblasts [Fig. 5(a)]. By combining the observation in Fig. 5(a) with the results in Fig. 4, we hypothesized that NTD treatment reduced the motility of fibroblasts in fibrin/Matrigel effectively, even without CDDP. Therefore, the suppression of cancer cell invasion by NTD might be mostly caused by the suppression of fibroblast activities.As another assessment of fibroblasts' activities, we measured the decrease in hydrogel areas caused by the fibroblast contraction4848. S. E. Park, A. Georgescu, J. M. Oh, K. W. Kwon, and D. Huh, ACS Appl. Mater. Interfaces 11, 23919–23925 (2019). https://doi.org/10.1021/acsami.9b07912 with and without NTD or PFD treatment. Figure 5(b) shows that without anti-fibrotic drugs, both MRC-5 fibroblasts (after 48 h of culture) and NHLFs (after 120 h of culture) in the A549 cell-conditioned medium induced area shrinkage of fibrin/Matrigel, but the contraction caused by NHLFs (∼20%) was much smaller than that caused by MRC-5 fibroblasts (∼50%). The 0.5 μM NTD treatment impeded the area shrinkage of fibroblast-laden fibrin/Matrigel. In contrast, the 1000-μM PFD treatment did not exhibit a similar effect. This result suggests that the contraction of fibroblasts in fibrin/Matrigel might be suppressed by NTD but not by PFD.The shrinkage of hydrogels caused by fibroblasts inside could lead to the underestimation of the growth and invasion of cancer cells. Therefore, with the co-culture of fibroblasts, the tumor growth and cancer cell invasion of the control and PFD-treated groups could be higher than what we measured. In contrast, for the NTD-treated group, the evaluations of spheroid areas and invasiveness should be more accurate. In other words, the efficacy of NTD in the suppression of tumor growth and cancer cell invasion in the co-culture system could be more evident than what we observed in Figs. 4(e)–4(g).To further determine the effects of anti-fibrotic drugs on fibroblasts in 3D culture, we performed RNA sequencing on MRC-5 fibroblasts in fibrin/Matrigel treated with 0.5 μM NTD or 1000 μM PFD for 48 h. The fibroblasts were in the medium conditioned by A549 cells (2 × 105 ml−1, 24 h of culture) to mimic the co-culture condition. We identified the expression of four genes relevant to cell adhesion, invasion, or ECM degradation that were reduced by NTD but increased or unaffected by PFD. The target genes were selected with two criteria: (1) the log2 (fold change between NTD and control groups) value was smaller than –0.42, which meant that NTD treatment reduced this gene expression by more than 25% compared with that in the control group, and (2) the log2(fold change between NTD and PFD groups) value was smaller than −1.0, which meant that this gene expression under NTD treatment was less than 50% compared with that under PFD treatment. The log2 values of fold changes between NTD and control groups, and between NTD and PFD groups, of these four genes (AFAP1L2, MMP16, ITGA8, and POSTN) are listed in Table I. As a comparison, the log2(fold change between NTD and PFD groups) values of common genes in CAFs, ACTA2 (α-smooth muscle actin), COL1A2 (collagen I), FN1 (fibronectin), and VIM (vimentin) were between –0.29 and 0.82, indicating only small differences in the expression of these genes between the fibroblasts treated by NTD and those by PFD.

TABLE I. Four genes relevant to cell adhesion, invasion, or ECM degradation in MRC-5 fibroblasts cultured in fibrin/Matrigel, whose expression were reduced by NTD treatment for more than 25% compared with those in the control group. The expression of these genes under NTD treatment was less than 50% compared with that under PFD treatment.

Gene symbolProtein descriptionLog2 (fold change between NTD and control groups)Log2 (fold change between NTD and PFD groups)AFAP1L2 (XB130)Actin filament-associated protein 1 like 2−1.57−1.37MMP16Matrix metallopeptidase 16−1.30−3.35ITGA8Integrin subunit alpha 8−1.08−2.85POSTNPeriostin−0.55−3.55Among the proteins encoded by the four genes listed in Table I, actin filament-associated protein 1 like 2 (encoded by AFAP1L2, also known as XB130) is known to be involved in the formation of lamellipodia, and the downregulation of AFAP1L2 could reduce cell migration and invasion.4949. M. Lodyga, X.-H. Bai, A. Kapus, and M. Liu, J. Cell Sci. 123, 4156–4169 (2010). https://doi.org/10.1242/jcs.071050 The reduced expression of MMP16 by NTD could lower the invasion capability into the ECM.5050. S. Quintero-Fabián, R. Arreola, E. Becerril-Villanueva, J. C. Torres-Romero, V. Arana-Argáez, J. Lara-Riegos, M. A. Ramírez-Camacho, and M. E. Alvarez-Sánchez, Front. Oncol. 9, 1370 (2019). https://doi.org/10.3389/fonc.2019.01370 The reduction of ITGA8 expression could decrease cell adhesion on ECM,5151. L. M. Schnapp, N. Hatch, D. M. Ramos, I. V. Klimanskaya, D. Sheppard, and R. Pytela, J. Biol. Chem. 270, 23196–23202 (1995). https://doi.org/10.1074/jbc.270.39.23196 and therefore, the fibroblasts were not able to extend into the spindle shape in the fibrin/Matrigel [Fig. 5(a)]. Meanwhile, decreased cell adhesion in the ECM could also reduce the contraction force on the hydrogel exerted by fibroblasts, which is in line with the results shown in Fig. 5(b). Finally, NTD could also reduce periostin (encoded by POSTN), a cell-adhesion protein that is highly expressed in various types of cancers and positively correlated with cancer invasion and metastasis.5252. A. Y. Liu, H. Zheng, and G. Ouyang, Matrix Biol. 37, 150–156 (2014). https://doi.org/10.1016/j.matbio.2014.04.007 Periostin secreted by activated fibroblasts in tissues of idiopathic pulmonary fibrosis could increase the proliferation of lung cancer cells.5353. H. Yamato, K. Kimura, E. Fukui, T. Kanou, N. Ose, S. Funaki, M. Minami, and Y. Shintani, Sci. Rep. 11, 21114 (2021). https://doi.org/10.1038/s41598-021-00717-5

Although to explore the detailed signaling pathways of proteins encoded by these genes under NTD or PFD treatment will require several additional experiments, RNA sequencing data suggest that the effects of NTD on CAFs could be more suppressive than PFD in terms of cell adhesion, invasion, or ECM degradation. This is consistent with the results of our 3D co-culture experiments. In addition, because NTD could reduce the expression of genes related to cell adhesion or migration, such as ITGA8 and AFAP1L2, and those genes controlling secretive proteins, such as MMP16 and POSTN, the physical and biochemical mechanisms must be taken into account together to correctly explain the effects of anti-fibrotic drugs on the cancer cell–fibroblast co-culture system.

NTD has been confirmed to improve the effects of anticancer drugs. For example, NTD combined with docetaxel was suggested to be second-line therapy for patients with non-small-cell lung cancer who have been treated with platinum-based drugs.5454. M. Reck, R. Kaise, A. Mellemgaard, J.-Y. Douillard, S. Orlov, M. Krzakowski, J. V. Pawel, M. Gottfried, I. Bondarenko, M. Liao, C.-N. Gann, J. Barrueco, B. Gaschler-Markefski, and S. Novello, Lancet Oncol. 15, 143–155 (2014). https://doi.org/10.1016/S1470-2045(13)70586-2 Recently, NTD combined with carboplatin plus nab-paclitaxel was found to improve the overall survival in lung cancer patients with idiopathic pulmonary fibrosis.5555. K. Otsubo, J. Kishimoto, M. Ando, H. Kenmotsu, Y. Minegishi, H. Horinouchi, T. Kato, E. Ichihara, M. Kondo, S. Atagi, M. Tamiya, S. Ikeda, T. Harada, S. Takemoto, H. Hayashi, K. Nakatomi, Y. Kimura, Y. Kondoh, M. Kusumoto, K. Ichikado, N. Yamamoto, K. Nakagawa, Y. Nakanishi, and I. Okamoto, Eur. Respir. J. 60, 2200380 (2022). https://doi.org/10.1183/13993003.00380-2022 In an animal study, NTD and an anti-VEGF/Ang2 nanobody were used together to reduce the brain metastases of lung cancer cells.5656. B. Kovalchuk, A. S. Berghoff, M. A. Karreman, K. Frey, M. Piechutta, M. Fischer, J. Grosch, S. Heiland, M. O. Breckwoldt, F. Hilberg, W. Wick, and F. Winkler, Clin. Exp. Metastasis 37, 637–648 (2020). https://doi.org/10.1007/s10585-020-10055-x The results of these clinical trials and the in vivo test are in agreement with that of our 3D co-culture experiments, supporting that NTD could improve the efficacy of anticancer drugs. On the other hand, we have not found any published results of clinical trials about the combined effect of PFD and anticancer drugs on lung cancer. Animal studies showed that PFD could improve the antitumor effects of cisplatin4141. M. Mediavilla-Varela, K. Boateng, D. Noyes, and S. J. Antonia, BMC Cancer 16, 176 (2016). https://doi.org/10.1186/s12885-016-2162-z or carboplatin.5757. A. Fujiwara, Y. Shintani, S. Funaki, T. Kawamura, T. Kimura, M. Minami, and M. Okumura, Lung Cancer 106, 8–16 (2017). https://doi.org/10.1016/j.lungcan.2017.01.006 However, the dose of PFD used in these two works was 200 mg/kg body weight per day, which is about five times that recommended for human patients.5858. S. Gulati and T. R. Luckhardt, Drug, Healthcare Patient Saf. 12, 85–94 (2020). https://doi.org/10.2147/DHPS.S224007 In a follow-up work, Fujiwara et al. demonstrated that PFD of the same dose alone could reduce the subcutaneous tumor volume of A549 cells.4646. A. Fujiwara, S. Funaki, E. Fukui, K. Kimura, T. Kanou, N. Ose, M. Minami, and Y. Shintani, Sci. Rep. 10, 10900 (2020). https://doi.org/10.1038/s41598-020-67904-8 Increasing the daily dose to 250 or 500 mg/kg, Marwitz et al. also showed that PFD reduced the tumor growth of subcutaneously transplanted Lewis lung carcinoma cells.5959. S. Marwitz, K. Turkowski, D. Nitschkowski, A. Weigert, J. Brandenburg, N. Reiling, M. Thomas, M. Reck, D. Drömann, W. Seeger, K. F. Rabe, R. Savai, and T. Goldmann, Front. Oncol. 9, 1550 (2020). https://doi.org/10.3389/fonc.2019.01550 The efficacy of high-dose PFD in animal studies should be valid, but the dosage might be much higher than that clinically applicable and tolerable in humans. In contrast, with the 3D co-culture model, we can control the drug concentrations more precisely, and therefore, the results could be more relevant to future clinical applications.