Dalpiciclib and Pyrotinib Exert Synergistic Antitumor Effects in Triple Positive Breast Cancer

Background The therapeutic benefit of the standard combination of anti-HER2 and chemotherapy in triple-positive breast cancer (TPBC) is limited even after the addition of endocrine therapy to the regimen. Therefore, treatment optimization is required urgently. Methods Through the drug sensitivity test, the drug combination efficacy of anti-HER2 drug, endocrine drug and CDK4/6 inhibitor to BT474 cells were tested. The underlying molecular mechanisms were investigated using immunofluorescence, western blot analysis, immunohistochemical staining and cell cycle analysis. Potential biomarker which may indicate the responsiveness to drug treatment in triple positive breast cancer was selected out using RNA-sequence and tested using immunohistochemical staining. Results We found that pyrotinib combined with dalpiciclib showed better efficacy than pyrotinib combined with tamoxifen in BT474 cells. Degradation of HER2 could enhance ER nuclear transportation, whereas cell cycle blockers could reverse this process. This may be the underlying mechanism by which the addition of dalpiciclib was more beneficial than the addition of pyrotinib plus tamoxifen. Furthermore, CALML5 was revealed to be a potential indicator of responsiveness to anti-HER2 therapy plus CDK4/6 inhibition in triple positive breast cancer. Conclusion Our study provided evidence for the introduction of CDK4/6 inhibitor in the treatment of TPBC and indicated that the combination of anti-HER2 therapy and cell cycle blockers may be a better strategy for TPBC treatment. Funding This study was supported by the National Natural Science Foundation of China (#U20A20381, #81872159)


Introduction 38
Human epidermal growth factor receptor 2-positive (HER2 + ) breast cancer is 39 associated with an increased risk of disease recurrence and death (Perou et al., 2000;40 Slamon et al., 1987;Tzahar et al., 1996). HER2-overexpressing breast tumors have 41 high heterogeneity, accounting partially for the co-expression of hormone receptors 42 (HR) (Loi et al., 2016). Previous studies have demonstrated that extensive cross-talk 43 exists between the HER2 signaling pathway and the estrogen receptor (ER) pathway 44 (Wang et al., 2011). In addition, exposure to anti-HER2 therapy may reactivate the ER 45 signaling pathway, which could lead to drug resistance (Brandao et al., 2020). 46 Generally, however, HER2-positive patients are treated using the same algorithms, 47 both in the early and advanced stages (Moja et al., 2012). Thus, novel therapeutic 48 strategies are urgently needed for patients with HER2 + /HR + breast cancer. 49 Increasing evidence has confirmed that the intrinsic differences between 50 HER2 + /HR + and HER2 + /HRpatients should not be ignored (Carey et al., 2016). 51 Clinical outcomes have demonstrated that HER2 + /HR + breast cancer patients have a 52 lower chance of achieving a pathologically complete response than HER2 + /HR -53 patients, when treated with neoadjuvant chemotherapy plus anti-HER2 therapy 54 (Cameron et al., 2017;Cortazar et al., 2014). Nevertheless, the addition of 55 concomitant endocrine therapy to anti-HER2 therapy or chemotherapy did not show 56 any advantages in clinical trials, such as the NSABP B-52 and ADAPT HER2 + /HR + 57 studies (Harbeck et al., 2017;Rimawi et al., 2017). Therefore, whether endocrine 58 therapy is useful in HER2 + /HR + breast cancer treatment is questionable. 59 4 efficacy in HER2 + breast cancer cells (Goel et al., 2016;Zhang et al., 2019). The 63 combination of CDK4/6 inhibitors and HER2-targeted therapy as an alternative 64 strategy for HER2 + /HR + patients, warrants further exploration. 65 Herein, we investigated the combined effect of pyrotinib (anti-HER2 drug), 66 tamoxifen (endocrine therapy), and dalpiciclib (CDK4/6 inhibitor) on the triple-67 positive breast cancer (TPBC) cell line BT474. We found that pyrotinib combined 68 with dalpiciclib showed better efficacy than pyrotinib combined with tamoxifen. In 69 addition, HER2-targeted therapy induced nuclear ER redistribution in TPBC cells,70 which could be reversed by the addition of a CDK4/6 inhibitor. Furthermore, 71 CALML5 could be a potential indicator of responsiveness to HER2-targeted therapy 72 combined with a CDK4/6 inhibitor. Our study provided a new function of the CDK4/6 73 inhibitor in TPBC cells treated with anti-HER2 therapy and suggested a novel strategy 74 for improving the clinical response in TPBC. 75

Results 76
Pyrotinib combined with dalpiciclib shows better efficacy than when combined 77 with tamoxifen 78 To explore the effects of pyrotinib, tamoxifen, and dalpiciclib in TPBC, we first 79 evaluated the cytotoxic activities of these three reagents in BT474 breast cancer cells. 80 The results showed that the IC50 doses for pyrotinib, tamoxifen, and dalpiciclib were 81 10 nM, 5 μM, and 8 μM, respectively ( Figure 1-figure supplement 1a). To further 82 investigate whether these three drugs could have a synergistic effect in BT474 cells, 83 we assessed the efficacies of the combinations of pyrotinib and dalpiciclib, pyrotinib 84 and tamoxifen, and tamoxifen and dalpiciclib on the inhibition of cell proliferation at 85 different concentrations. We calculated the combination index for each combination 86 using Compusyn software to determine if the antitumor effects were synergistic (Chou 87 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted October 4, 2021. ; https://doi.org/10.1101/2021.10.04.463019 doi: bioRxiv preprint 5 and Talalay, 1984). Synergistic effects were observed in the combination group of 88 pyrotinib and dalpiciclib, as well as in the pyrotinib and tamoxifen groups; both with 89 CI values of <1 at several concentrations ( Figure 1a). However, in the combination 90 group of tamoxifen and dalpiciclib, no synergistic effect was observed. 91 We also analyzed the effect of the three-drug combination, and it showed a 92 stronger cytotoxic effect on TPBC compared with the effect of the other two-drug 93 combinations ( Figure 1b). As both dalpiciclib and tamoxifen showed synergistic 94 effects in combination with pyrotinib, we sought the combination that showed better 95 efficacy. Hence, we treated the BT474 cells with different combinations at EC50 or 96 half EC50 concentrations. The three-drug combination and the combination of 97 pyrotinib and dalpiciclib showed a stronger cell inhibition compared with that exerted 98 by pyrotinib and tamoxifen as well as tamoxifen and dalpiciclib ( Figure 1c). The 99 colony formation assay also showed similar trends as the cell viability assay; the 100 three-drug combination formed the least number of colonies, followed by the 101 combination of pyrotinib and dalpiciclib ( Figure 1-figure supplement 1b-c). 102 To verify the efficacy of endocrine therapy in HR + /HER2 + patients, we 103 retrospectively analyzed the clinical data of 221 HR + /HER2 + patients who received 104 adjuvant therapy at the Shengjing Hospital. Of these, 44 patients received anti-HER2 105 therapy plus chemotherapy, and 177 patients received anti-HER2 therapy combined 106 with chemotherapy and endocrine therapy. A Kaplan-Meier analysis showed that the 107 addition of endocrine therapy to adjuvant anti-HER2 therapy plus chemotherapy did 108 not significantly alter disease-free survival (DFS; P = 0.600) or overall survival (OS; 109 P = 0.5276) in HR + /HER2 + patients ( Figure 1d). 110 111 Nuclear ER distribution is increased after Anti-HER2 therapy and could be 112 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted October 4, 2021. ; https://doi.org/10.1101/2021.10.04.463019 doi: bioRxiv preprint 6 reversed by the introduction of a CDK4/6 inhibitor 113 The results of the drug sensitivity test showed that the combination of pyrotinib 114 and tamoxifen was less effective than the combination of pyrotinib and dalpiciclib. 115 Considering that anti-HER2 therapy may activate the ER signaling pathway, we 116 performed immunofluorescence staining for ER distribution on the different drug-117 treated groups. We found that pyrotinib induced ER nuclear translocation in BT474 118 cells, which could be partially reversed by the addition of dalpiciclib, rather than 119 tamoxifen ( Figure 2a). To further investigate whether the expression of HER2 could 120 affect the distribution of ER, we transfected MCF7 cells with HER2 overexpression 121 plasmids. We found that ER in MCF7 cells (WT and NC) was mainly expressed in the 122 nuclei, whereas in the HER2 overexpressing MCF7 cells, ER was distributed 123 throughout the cell plasma. Treating HER2 overexpression MCF7 cells with pyrotinib 124 could redistribute the ER into the nuclei (Figure 2b). Western blot analyses revealed 125 although the nuclear ER levels increased considerably, the total expression of ER was 126  (Table 1). We found significant elevations in the nuclear ER 134 expression levels of patients who received chemotherapy and anti-HER2 therapy, 135 compared with the levels in patients who only received chemotherapy (Figure 2c, d). 136 However, in our ongoing clinical trial (NCT04486911), the nuclear ER expression 137 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted October 4, 2021. The inhibition of the ER pathway might be involved in the effect of pyrotinib plus 164 dalpiciclib on TPBC cells; therefore, intersection analyses were performed to confirm 165 this. As shown in Figure 3g, CALML5, KRT15, and KRT19 are the common genes 166 shared between the two sets, the upregulated genes treated with pyrotinib and the 167 genes belonging to the estrogen signaling pathway. Since dalpiciclib is a cell cycle 168 blocker, we also analyzed the common genes involved in the upregulation of the 169 genes and the cell cycle progression after pyrotinib treatment. CDKN1A was the only 170 shared gene in these two sets ( Figure 3h). We then investigated whether any of the 171 above-mentioned genes were upregulated with the use of pyrotinib and whether this 172 could be reversed with the introduction of dalpiciclib, which may serve as a potential 173 biomarker of the responsiveness to different treatments. The results showed that only 174 one factor, CALML5, was the common gene ( Figure 3i). 175 CALML5 is a potential indicator for the responsiveness to anti-HER2 therapy 176 plus CDK4/6 inhibitor 177 Western blot analyses and bioinformatic analyses were conducted to verify the 178 changes in the signaling pathways. The western blot analyses showed that while the 179 introduction of tamoxifen did not significantly affect the expression of HER2 and 180 partially inhibited the HER2 downstream pathway (AKT-mTOR signaling pathway), 181 it did not significantly affect the phosphorylation of Rb (Figure 4a). In contrast, after 182 the introduction of dalpiciclib, the activation of mTOR was partially inhibited, which 183 relieved the negative feedback on the HER2 pathway, as evidenced by the slight 184 increase in the HER2 and pAKT, which maintained the sensitivity of the HER2 185 pathway to pyrotinib. This was consistent with the findings of Goel et al (Goel et al., 186 2016). The combination of pyrotinib and dalpiciclib significantly reduced pRb 187 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted October 4, 2021. Until now, the combination of antiHER2 therapy and chemotherapy have been 205 the major treatment strategies for treatment of TPBC, although sometimes this is 206 combined with endocrine therapy (Gianni et al., 2012;Schneeweiss et al., 2013). 207 Although pCR and DFS improve with the use of the combination of anti-HER2 208 therapy and chemotherapy, the strong adverse effects of chemotherapy cannot be 209 ignored (Maguire et al., 2021). Moreover, clinical data showed that the addition of 210 anti-estrogen receptor drugs in the treatment regimen of TPBC did not provide 211 additional advantages in the pCR rates and DFS (Harbeck et al., 2017;Rimawi et al., 212 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted October 4, 2021. ; https://doi.org/10.1101/2021.10.04.463019 doi: bioRxiv preprint kinase inhibitors (TKIs) and CDK4/6 inhibitors, additional chemo-free strategies are 214 being developed for the treatment of HER2 + /HR + breast cancer (Gianni et al., 2018;215 the addition of CDK4/6 inhibitors to HER2-targeted and endocrine therapy could be 217 an alternative strategy for HER2 + /HR + patients. 218 In our study, we found that the combination of tamoxifen and pyrotinib was less 219 effective than the combination of pyrotinib and dalpiciclib in BT474 cancer cells. This 220 was anomalous since the two blocking agents of HER2 and ER were expected to 221 inhibit their crosstalk and achieve better responses. To explore the potential 222 mechanisms, we investigated the crosstalk between HER2 and the ER. After 223 degrading HER2 with pyrotinib, ER was found to relocate to the cell nucleus, 224 enhancing the function of ER which was consistent with the findings of Kumar et al 225 and Yang et al (Kumar et al., 2002;Yang et al., 2004). We believe that the anti-HER2 226 mediated ER redistribution caused the enhanced ER function, leading to the relatively 227 low efficacy of the combination of pyrotinib and tamoxifen in the treatment of 228 HER2 + /HR + cells. Moreover, we found that the introduction of dalpiciclib to pyrotinib 229 significantly decreased the total and nuclear expression of ER, reversing the ER 230 activation caused by pyrotinib (Figure2-figure supplement 1c). This may be the 231 underlying mechanism by which the addition of a CDK4/6 inhibitor was more 232 beneficial than the addition of pyrotinib and tamoxifen. 233 Furthermore, using RNA-seq and bioinformatics analyses, CALML5 was 234 selected as a potential biomarker for responsiveness to anti-HER2 therapy combined 235 with a CDK4/6 inhibitor. CALML5, known as calmodulin-like 5, is a skin-specific 236 calcium-binding protein that is closely related to keratinocyte differentiation (Mehul 237 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted October 4, 2021. ; https://doi.org/10. 1101/2021.10.04.463019 doi: bioRxiv preprint et al., 2001. A previous study showed that the high expression of CALML5 was 238 strongly associated with better survival in patients with head and neck squamous cell 239 carcinomas (Wirsing et al., 2021). Misawa et al. (Misawa et al., 2020) reported that 240 the methylation of CALML5, led to its downregulation, and this showed a correlation 241 with HPV-associated oropharyngeal cancer. Moreover, the ubiquitination of 242 CALML5 in the nucleus was found to play a role in the carcinogenesis of breast 243 cancer in premenopausal women (Debald et al., 2013). Our results suggested that 244 TPBC patients with positive CALML5 may benefit from the addition of CDK4/6 245 inhibitors in neoadjuvant therapy. However, the underlying mechanism of CALML5 246 in breast cancer requires further investigation. 247 In conclusion, our study showed the novel role of the CDK4/6 inhibitor in TPBC 248 and provided evidence that CALML5 may be a potential biomarker in the prediction 249 of the responsiveness of HER2 + /HR + breast cancer patients to CDK4/6 inhibitors. 250 251 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted October 4, 2021. (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted October 4, 2021. ; https://doi.org/10.1101/2021.10.04.463019 doi: bioRxiv preprint

Colony formation assays 302
Cells were seeded in 6-well plates at a density of 1000 cells/well. The cells were 303 treated with DMSO (0.1%), pyrotinib (10 nM), tamoxifen (5 μM), dalpiciclib (8 μM), 304 or a combination of the two or three agents. During the process, the culture medium 305 was renewed every three days. After 14 days, the colonies were fixed and stained with 306 crystal violet. Clusters of more than eight cells were counted as colonies. This assay 307 was performed in triplicates independently. 308 Transfection of the human HER2 plasmids 309 MCF7 cells were cultured in DMEM supplemented with 10% FBS. Human 310 HER2 plasmids were purchased from Hanbio. The plasmids were mixed well with 311 lipo3000 and p3000, according to the manufacturer's instructions, and then added to 312 the culture medium. Forty-eight hours after the transfection, the cells were fixed in 313 formaldehyde and stained for the estrogen receptor. 314

Western blot analysis 315
Cells and cancer tissues were lysed using a cell lysis buffer (Beyotime,Shanghai,316 China). The total proteins were extracted in a lysis buffer (Beyotime,Shanghai,317 China), and the nuclear proteins were extracted using a nuclear protein extraction kit 318 (Beyotime), in which PMSF, protease, and phosphatase inhibitors were added. Protein 319 concentrations were determined using a Pierce BCA Protein Assay Kit (Thermo 320 Fisher Scientific, Waltham, MA, USA) according to the manufacturer's instructions. 321 The proteins from the cells and tissue lysates were separated using 10% SDS-PAGE 322 and 6% SDS-PAGE, respectively, and then transferred to polyvinylidene fluoride 323 (PVDF) membranes. The immunoreactive bands were detected using enhanced 324 chemiluminescence (ECL). The western blot analysis was performed in triplicates 325 independently. 326 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted October 4, 2021. ; https://doi.org/10.1101/2021.10.04.463019 doi: bioRxiv preprint

Immunofluorescence assays 327
The cellular localization of different proteins was detected using 328 immunofluorescence. Briefly, the cells grown on glass coverslips were fixed in 4% 329 paraformaldehyde at room temperature for 30 min. Cells were incubated with the 330 respective primary antibodies for 1 h at room temperature, washed in PBS, and then 331 incubated with 590-Alexa-(red) secondary antibodies (Molecular Probes, Eugene, 332 OR, USA). We used 590-Alexa-phalloidin to localize the ER. The nuclei of the cells 333 were stained with DAPI and color-coded in blue. The images were captured using an 334 immunofluorescence microscope (Nikon Oplenic Lumicite 9000). The distribution 335 ratio of ER was calculated manually by randomly chosen 5 views in 336 400magnification. The immunofluorescence assay was performed in triplicates 337 independently. 338

Immunohistochemical staining 339
The clinical samples were fixed in 4% formaldehyde, embedded in paraffin, and 340 sectioned continuously at a thickness of 3 μm. The paraffin sections were 341 deparaffinized with xylene and rehydrated using a graded ethanol series. They were 342 then washed with tris-buffered saline (TBS). After these preparation procedures, the 343 sections of each sample were incubated with the primary anti-ER antibody (Abcam 344 Company, ab32063), anti-HER2 antibody (Abcam Company, ab134182), and anti-345 CALML5 antibody (Proteintech, 13059-1-AP) at 4 °C overnight. The next day, they 346 were washed three times with TBS and incubated with a horseradish peroxidase 347 (HRP)-conjugated secondary antibody (Gene Tech Co. Ltd., Shanghai, China) at 348 37 °C for 45 min, followed by immunohistochemical staining using a DAB kit (Gene 349 Tech Co. Ltd.) for 5-10 min. 350

Evaluation of the ER and HER2 statuses 351
. CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted October 4, 2021. ; https://doi.org/10.1101/2021.10.04.463019 doi: bioRxiv preprint

16
The ER and HER2 statuses of patients who received neoadjuvant therapy were 352 evaluated by a pathologist from a Shenjing affiliated hospital. The clinical specimens 353 before and after the neoadjuvant therapy were evaluated. The analyses of the elevation 354 or decline in ER statuses were based on these pathological reports. 355

Gene enrichment analysis 356
Gene annotation data in the GO and KEGG databases and R language were used 357 for the enrichment analysis. Only enrichment with q-values less than 0.05 were 358 considered significant. 359

GSEA 360
The hallmark gene sets in the Molecular Signatures Database were used for 361 performing the GSEA; only gene sets with q-values less than 0.05 were considered 362 significantly enriched. 363

Statistical analysis 364
All the descriptive statistics (except the drug sensitivity assay in Figure 1a and b) 365 were presented as the means ± standard deviations (SDs). The drug sensitivity assay 366 in Figure 1 a and b were presented as the means ± standard error of mean (SEM). The 367 differences between the groups were analyzed by chi-squared or Student's t tests. 368 Kaplan-Meier methods were used to compute the survival analysis and P-value was 369 obtained by log-rank test. The statistical analyses were performed using IBM SPSS 370 version 22 (SPSS, Armonk, NY, USA) and GraphPad Prism version 7. The statistical 371 significance of the differences between the test and control samples was assessed at 372 significance thresholds of *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. 373 374 . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted October 4, 2021. ;https://doi.org/10.1101https://doi.org/10. /2021  . CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted October 4, 2021. ;https://doi.org/10.1101https://doi.org/10. /2021   CC-BY 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted October 4, 2021. ; https://doi.org/10.1101/2021.10.04.463019 doi: bioRxiv preprint