English

• Mass spectrometry imaging (MSI)-based spatial metabolomics is a powerful technology for visualizing the distribution and expression of functional molecules in disease-associated metabolic phenotypes [1-3]. This technology is particularly valuable for understanding the metabolic heterogeneity in tumor tissues [4-6], kidneys [7], hearts [8-11], brains [12], and other organs [13-16]. Transcriptomics can explore the spatiotemporal dynamics of cell types and gene expression as a reflection of changes in biological tissues [17, 18]. Multi-omics analysis allows systematic interrogation of up and downstream molecular information, linking differential metabolites, proteins, and genes, thereby enabling a more precise characterization of molecular changes in heterogeneous tissues [19-21].

  After the onset of heart failure (HF), the heart undergoes notable structural changes, including ventricular dilation, thinning of the ventricular wall, and even rupture, often leading to the formation of distinct areas of pathological damage. This progression is closely accompanied by spatiotemporal alterations in cellular and molecular activity [22, 23]. Spatial omics analyses have been applied to in situ studies of the heterogeneous changes among different microareas in the heart after the onset of heart diseases [24, 25]. For example, spatial metabolomics revealed changes in endogenous metabolites and spatial heterogeneity across heart chambers in diabetic cardiomyopathy [26]. Spatial transcriptome studies have shown that the infarct margin (IM) may manifest unique responses following the onset of HF [9, 27]. However, in situ exploration of the molecular characteristics of HF-associated heart tissue and drug interventions based on pathological considerations remains scarce. Developing a new method to comprehensively characterize HF-related molecular changes in different microareas of heart tissue with HF is crucial for understanding the pathophysiology of HF and identifying potential therapeutic targets.

  Here, we integrated spatial metabolomics and transcriptomics technologies to delineate the molecular changes in metabolites and genes in the heart tissue of rats with HF. We then used the same strategy to assess the therapeutic properties of two clinical HF drugs, valsartan and Qishen Yiqi Dripping Pills (QDP), and identify putative therapeutic targets associated with drug efficacy. The research workflow is shown in Fig. S1 (Supporting information).

  Male adult Sprague-Dawley rats weighing 200–250 g and aged 6–7 weeks were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. All animal experimental procedures were conducted following the regulations of the Animal Experimental Management and Animal Welfare Ethics Committee of the Institute of Materia Medica, Chinese Academy of Medical Sciences and Peking Union Medical College (Approval No. 00004235). The rats were randomly divided into four groups (n = 15): sham surgery (S), model (M), valsartan administration (V), and QDP administration (Q). Following four weeks of drug administration, rat heart tissues were collected, frozen in liquid nitrogen, and stored at −80 ℃ for histopathological staining and spatial metabolomics research. In addition, 100 mg of tissue from the heterogeneous areas below the ligature were separated, washed with physiological saline, and stored at −80 ℃ in 1.5 mL tubes for transcriptomic analysis. Please refer to the Supporting information for detailed information on the animal experiments, sample pretreatment, and analysis processes.

  The rat model of cardiac insufficiency was established by ligating the left anterior descending coronary artery (Fig. S2 in Supporting information). H & E and Masson staining showed that the left ventricular cardiac tissue was partitioned into three distinct areas: the infarct (I), IM, and non-infarct (NI) areas (Figs. S3 and S4 in Supporting information). The spatial distribution of various metabolites in heart tissue was visualized using airflow-assisted desorption electrospray ionization (AFADESI)-MSI. Representative MS images are shown in Fig. S5 (Supporting information). The spatial resolution of AFADESI-MSI allowed for the clear division of heart tissue into multiple microareas. For example, the metabolite ion m/z 313.2353 was exclusively distributed in the I area, while metabolite ion m/z 832.5823 was concentrated in the IM area, and metabolite ion m/z 258.1550 was distributed in the NI area. An overlay image showed that the heart tissue of rats with HF could be accurately divided into three microareas along the coronal plane, which was consistent with the results of H & E and Masson staining (Fig. 1A). The significant differences in the mass spectrum between the I, IM, and NI areas were apparent (Figs. 1B-D), suggesting clear metabolic heterogeneity.

  Figure 1

  

  Figure 1.  Spatial profiling of heterogeneous metabolic characteristics in different microareas of the heart tissue in rats with HF using AFADESI-MSI. (A) MS images of (1) m/z 313.2353, (2) m/z 832.5823, (3) m/z 258.1550, (4) Overlay image of three ions, (5) H & E, and (6) Masson staining in the heart tissue of rats with HF. Mass spectrum of (B) I area, (C) IM area, and (D) NI area in the same tissue slice. I: Infarct. IM: Infarct margin. NI: Non-infarct.

  DownLoad: Full-Size Img PowerPoint

  Multivariate analysis showed that metabolites in energy metabolism–related pathways exhibited differentiation and clustering trends between the I, IM, and NI areas (Fig. S6 in Supporting information). Transcriptome analysis by Bulk RNA-seq was used to establish up and downstream associations of metabolites and genes in the I and NI areas. Significant transcriptional differences were observed between the M and the S groups in both I and NI areas (Figs. S7A-G in Supporting information). We integrated the differential metabolites and genes in a correlation analysis (Fig. S8 in Supporting information) and constructed a metabolome–transcriptome correlation network related to energy metabolism substrates. This network revealed substantial alterations in energy metabolism in the hearts of rats with HF, including the tricarboxylic acid (TCA) cycle, amino acid metabolism, and fatty acid (FA) metabolism (Fig. S7H in Supporting information).

  Among the TCA cycle-related metabolites, the relative intensities of glucose, lactic acid, citric acid, and malic acid increased significantly in the NI area of heart tissue in the M group, while fumaric acid and succinic acid decreased significantly (Fig. 2A). Bulk RNA-seq results also showed metabolic disorders in the TCA cycle in the NI area of the M group (Fig. S9 and Table S1 in Supporting information). The mRNA expression of PDHB, LDHA, and TCA-related genes FH, SDHA, MDH1, ACO2, and GOT1 were downregulated (Fig. 2B). Furthermore, ACLY increases cytosolic acetyl-CoA through citrate cleavage, linking FA synthesis and glucose metabolism [28, 29], and was significantly upregulated in the I and NI areas (Fig. 2B). These results indicated that the TCA cycle in NI areas was abnormally obstructed, and NI areas exhibited a non-standard citric acid–malic acid shuttle cycle (Fig. 2C) [30]. Most of these metabolites and genes were nearly completely depleted in the I area, and there were no statistically significant differences in the IM area (Figs. 2A and B).

  Figure 2

  

  Figure 2.  Reprogrammed TCA cycle in the heart tissue of rats with HF. (A) AFADESI-MS images of key metabolites in the heart tissues. (B) Violin plots of key TCA gene expression in the I and NI areas of the heart tissues between the S and M groups. (C) Metabolic pathway of the TCA cycle. Compared with the S group, red represents an increase in the NI area of the M group, and blue represents a decrease. n = 3, x ± SD. ^*P < 0.05, ^**P < 0.01. vs. S/M. S: sham surgery. M: model.

  DownLoad: Full-Size Img PowerPoint

  Owing to the disturbance in glucose metabolism, the heart utilizes amino acids as substrates for energy production [31]. The heart exhibited disordered amino acid metabolism (Fig. S10, Tables S2 and S3 in Supporting information). There was a decrease in antioxidant amino acids such as glutamate, glutamine, glutathione, hypotaurine, and taurine (Fig. S11A in Supporting information). Glutamate synthesis genes BCAT2 and ALDH5A1 were significantly downregulated (Fig. S11B in Supporting information), suggesting that oxidative stress was prevalent in the different damaged microareas of HF cardiac tissue and that the I area was significantly depleted of antioxidants [32]. In addition, differential metabolites in the pathway exhibited varying degrees of microarea specificity (Figs. S11 and S12 in Supporting information).

  FAs are the main energy source for myocardial cells and play a crucial role in the steady state of cardiac energy metabolism [33]. Carnitines act as transport mediators and play a key role in FA β-oxidation [34]. In the present study, carnitine, acetylcarnitine (CAR (2:0)), palmitoylcarnitine (CAR (16:0)), and stearoylcarnitine (CAR (18:0)) were reduced in the I area of the M group. Only the relative intensity of CAR (2:0) was significantly increased in the IM area, while in the NI area, the carnitines were significantly increased compared with the S group (Fig. S13A in Supporting information). The different distributions of carnitines in each injury microarea suggested significant differences in FA oxidation. Given the up–downstream relationship of FA oxidation, according to the gene IPA analysis results, carnitine metabolism and the mitochondrial FA β-oxidation pathway were significantly inhibited (Figs. S13B and S14, Table S1 in Supporting information), corresponding to the downregulation of CPTs and CRAT (Fig. S13C in Supporting information). Due to severe heart damage in the I area, the material basis was lost, regulatory genes were reduced, and the normal operation of the carnitine shuttle system was obstructed. In NI areas, carnitines significantly increased, accompanied by downregulated genes related to the carnitine shuttle system. In NI areas with less severe damage, due to the downregulation of regulatory genes, carnitines could not enter the carnitine shuttle normally and accumulated.

  To further explore the abnormal oxidation of FA, we conducted MSI analysis on the key metabolites of FA oxidation, namely oxo FA and fatty aldehydes (FALs). Oxo FAs and FALs are difficult to detect through MSI due to difficult ionization, low abundance, and other factors. In the present study, a hydrogel-assisted on-tissue chemical derivatization (OTCD) approach coupled with AFADESI-MSI was used to visualize oxo FAs and FALs (Fig. 3A) [12, 35-38]. As shown in Fig. 3B, after the occurrence of HF, FAL (18:0), FAL (18:1), and FAL (18:2) significantly increased in the NI area. FA (18:0), FA (18:1), and FA (18:2) decreased significantly in the NI area of the heart in the M group, while the oxo FA (18:1) increased significantly. With the OTCD, comprehensive abnormalities in FA metabolism were observed, suggesting disordered FA oxidation in the NI area of the heart. This dysfunction led to incomplete FA oxidation and a reduced energy supply to myocardial cells.

  Figure 3

  

  Figure 3.  Visualization of reprogrammed FA oxidative metabolism in the heart tissue of rats with HF using AFADESI-MSI after OTCD. (A) Schematic diagram of OTCD. (B) MS images of representative FALs, FAs and oxo FAs in the heart tissues from the S and M groups. n = 3, x ± SD. ^*P < 0.05, ^**P < 0.01 vs. S/M.

  DownLoad: Full-Size Img PowerPoint

  MSI enables in situ characterization of metabolic alterations in heterogeneous cardiac tissue after 4 weeks of valsartan and QDP administration. To our surprise, valsartan demonstrated a notable ability to improve the abnormal distribution of metabolites in the I area, while QDP had a more significant improvement effect on energy metabolites in the NI area (Fig. S15 in Supporting information). As shown in Fig. 4A and Fig. S16A (Supporting information), differential metabolites such as FA (4:0; O2), FAL (10:0), citric acid, malic acid, glycerol 3-phosphate (G3P), and uridine showed similar results. QDP increased the mRNA expression of CPT2, PDHB ACADVL, BCKDHB, and SUCLA2, and reduced the abnormally high mRNA expression of ACLY in the NI area, suggesting that QDP improved the NI area by regulating the TCA cycle as well as FA and branched-chain amino acid (BCAA) metabolism (Fig. 4B and Fig. S16B in Supporting information). In contrast, valsartan reversed the decrease in mRNA expression of CRAT, ACADVL, BCKDHB, and SUCLA2, optimized the distribution of FA and FAL, and restored FA and BCAA metabolism in the I area. We verified the proteins associated with these metabolic pathways by western blotting (Figs. S17 and S18 in Supporting information).

  Figure 4

  

  Figure 4.  Valsartan and QDP showed different metabolic regulation characteristics in heart tissue of HF rats. (A) MS images of representative metabolites in groups S, M, Q, and V, and ion intensity bar charts corresponding to different damaged microareas. (B) Violin plots of CPT2, CRAT, PDHB, and ACLY expression in the I and NI areas. n = 3, x ± SD. P < 0.05, ^**P < 0.01 vs. S/M. ^#P < 0.05, ^##P < 0.01 vs. Q/M, V/M.

  DownLoad: Full-Size Img PowerPoint

  In conclusion, we developed a method that combined spatial metabolomics and transcriptomics to comprehensively elucidate the genetic and metabolic changes in heterogeneous microareas of rat heart tissue affected by HF. This method accurately depicted specific molecular changes in the I, IM, and NI areas, particularly highlighting changes in the TCA cycle, amino acid metabolism, and FA metabolism. Furthermore, four potential therapeutic targets (CPT1A, PDHB, ACLY, and BCAT2) were identified and preliminarily validated for the first time in HF. These findings not only reflected the pathological characteristics of HF but also provided valuable insights into the pharmacological effects of distinct therapeutic agents. The combination of spatial metabolomics and transcriptomics offers an efficient method for studying tissue heterogeneity and drug research.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  CRediT authorship contribution statement

  Yue Xu: Writing – original draft, Investigation. Lingzhi Wang: Writing – review & editing. Liu Yang: Investigation. Renliang Xue: Validation. Haowen Zhu: Methodology. Qifeng Liu: Visualization. Xueqi Lv: Methodology. Ruiping Zhang: Project administration. Jun Tu: Validation, Data curation. Qingce Zang: Writing – review & editing, Project administration. Yinghong Wang: Writing – review & editing, Project administration.

  Acknowledgments

  This study was supported by the National Natural Science Foundation of China (No. 82374158), National Science and Technology Major Project (No. 2018ZX09711001–002–004), the Jiangxi University of Chinese Medicine Science and Technology Innovation Team Development Program (No. CXTD22007), and the Medical and Health Technology Innovation Project (No. 2022-I2M-1–020).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2025.110958.

  
  
• 1. [1]

     X. Song, C. Li, Y. Meng, Acta Mater. Medica 1 (2022) 507–533.

  2. [2]

     L. Li, Q. Zang, X. Li, et al., J. Pharm. Anal. 13 (2023) 483–493. doi: 10.1016/j.jpha.2023.04.007

  3. [3]

     Y. Chen, X. Ding, J. Zhou, et al., Chin. Chem. Lett. 36 (2025) 110351.

  4. [4]

     T. Risom, P. Chang, S. Rost, J. Ziai, Mass spectrometry-based tissue imaging of the tumor microenvironment, in: U.N. Kasid, R. Clarke (Eds.), Cancer Systems and Integrative Biology, Methods in Molecular Biology, New York, 2023, vol. 2660, pp. 171–185. 10.1007/978-1-0716-3163-8_12

  5. [5]

     C.L. Sun, A.Q. Wang, Y.H. Zhou, et al., Nat. Commun. 14 (2023) 2692.

  6. [6]

     W.C. Hsieh, B.R. Budiarto, Y.F. Wang, et al., J. Biomed. Sci. 29 (2022) 96. doi: 10.1186/s12929-022-00879-y

  7. [7]

     I. Abbas, M. Noun, D. Touboul, et al., Int. J. Mol. Sci. 20 (2019) 1623. doi: 10.3390/ijms20071623

  8. [8]

     Y.H. Liu, X. Zhang, S. Yang, et al., J. Pharm. Anal. 13 (2023) 1496–1509. doi: 10.1016/j.jpha.2023.08.011

  9. [9]

     K. Kanemaru, J. Cranley, D. Muraro, et al., Nature 619 (2023) 801–810. doi: 10.1038/s41586-023-06311-1

  10. [10]

      J. Ren, H.W. Li, L. Chen, et al., Research 6 (2023) 0019. doi: 10.34133/research.0019

  11. [11]

      R.F. Menger, W.L. Stutts, D.S. Anbukumar, et al., Anal. Chem. 84 (2012) 1117–1125. doi: 10.1021/ac202779h

  12. [12]

      L. Wang, Q. Zang, X. Tu, et al., Talanta 269 (2024) 125491.

  13. [13]

      P. Zheng, N. Zhang, D. Ren, et al., Cell Rep. Med. 4 (2023) 101057. doi: 10.1016/j.xcrm.2023.101057

  14. [14]

      P. Chen, Z. Zhu, H. Geng, et al., J. Pharm. Anal. 14 (2024) 100910. doi: 10.1016/j.jpha.2023.11.017

  15. [15]

      C. Zhao, P.S. Xie, T. Yang, et al., Chin. Chem. Lett. 29 (2018) 1281–1283.

  16. [16]

      K. Vandereyken, A. Sifrim, B. Thienpont, et al., Nat. Rev. Genet. 24 (2023) 494–515. doi: 10.1038/s41576-023-00580-2

  17. [17]

      A.L. Koenig, I. Shchukina, J. Amrute, et al., Nat. Cardiovasc. Res. 1 (2022) 263–280. doi: 10.1038/s44161-022-00028-6

  18. [18]

      M.E. Sweet, A. Cocciolo, D. Slavov, et al., BMC Genomics. 19 (2018) 812. doi: 10.1186/s12864-018-5213-9

  19. [19]

      X. Liu, M. Jin, M. Zhang, et al., J. Pharm. Biomed. Anal. 172 (2019) 126–138. doi: 10.1504/ijil.2019.097674

  20. [20]

      V.M. Ravi, P. Will, J. Kueckelhaus, et al., Cancer Cell 40 (2022) 639–655. doi: 10.1016/j.ccell.2022.05.009

  21. [21]

      E. Bertero, C. Maack, Nat. Rev. Cardiol. 15 (2018) 457–470. doi: 10.1038/s41569-018-0044-6

  22. [22]

      A. Gupta, B. Houston, Heart Fail Rev 22 (2017) 825–842. doi: 10.1007/s10741-017-9623-6

  23. [23]

      D.M. Calcagno, N. Taghdiri, V.K. Ninh, et al., Nat. Rev. Cardiol. 1 (2022) 1039–1055. doi: 10.1038/s44161-022-00160-3

  24. [24]

      S. Hayat, R. Kramann, Nat. Cardiovasc Res. 1 (2022) 978–979. doi: 10.1038/s44161-022-00161-2

  25. [25]

      D.A. Dean, G. Gautham, J.L. Siqueira-Neto, et al., PLoS Negl. Trop. Dis. 15 (2021) e0009819. doi: 10.1371/journal.pntd.0009819

  26. [26]

      H. Wu, X. Liu, Z.Y. Gao, et al., Int. J. Mol. Sci. 20 (2019) 4837. doi: 10.3390/ijms20194837

  27. [27]

      D.M. Calcagno, N. Taghdiri, V.K. Ninh, et al., Nat. Cardiovasc Res. 1 (2022) 1039–1055. doi: 10.1038/s44161-022-00160-3

  28. [28]

      X. Feng, L. Zhang, S. Xu, et al., Prog. Lipid Res. 77 (2020) 101006. doi: 10.1016/j.plipres.2019.101006

  29. [29]

      W. Xiang, H. Lv, F. Xing, et al., Sci. Adv. 9 (2023) eadi2465. doi: 10.1126/sciadv.adi2465

  30. [30]

      N.C. Williams, L.A.J. O'Neill, Front. Immunol. 9 (2018) 141. doi: 10.3389/fimmu.2018.00141

  31. [31]

      D. Murashige, C. Jang, M. Neinast, et al., Science 370 (2020) 364–368. doi: 10.1126/science.abc8861

  32. [32]

      K. Stathopoulou, I. Wittig, J. Heidler, et al., FASEB J 30 (2016) 1849–1864. doi: 10.1096/fj.201500048

  33. [33]

      S. Todisco, B. Musio, V. Pesce, et al., Biochem. Pharmacol. 208 (2023) 115405. doi: 10.1016/j.bcp.2022.115405

  34. [34]

      D.M. Muoio, P.D. Neufer, Cell Metab 15 (2012) 595–605. doi: 10.1016/j.cmet.2012.04.010

  35. [35]

      Y.Q. Cao, Y.J. Lu, L. Zhang, et al., Chin. Chem. Lett. 35 (2024) 109788.

  36. [36]

      Q. Zang, M. Wang, Y. Zhu, et al., Anal. Chem. 93 (2021) 15373–15380. doi: 10.1021/acs.analchem.1c03118

  37. [37]

      L. Wang, Q. Zang, Y. Zhu, et al., Anal. Chem. 95 (2023) 1975–1984. doi: 10.1021/acs.analchem.2c04316

  38. [38]

      J. Liu, Q. Zang, X. Li, et al., Chin. Chem. Lett. 34 (2023) 108322.

• 

  Figure 1  Spatial profiling of heterogeneous metabolic characteristics in different microareas of the heart tissue in rats with HF using AFADESI-MSI. (A) MS images of (1) m/z 313.2353, (2) m/z 832.5823, (3) m/z 258.1550, (4) Overlay image of three ions, (5) H & E, and (6) Masson staining in the heart tissue of rats with HF. Mass spectrum of (B) I area, (C) IM area, and (D) NI area in the same tissue slice. I: Infarct. IM: Infarct margin. NI: Non-infarct.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Reprogrammed TCA cycle in the heart tissue of rats with HF. (A) AFADESI-MS images of key metabolites in the heart tissues. (B) Violin plots of key TCA gene expression in the I and NI areas of the heart tissues between the S and M groups. (C) Metabolic pathway of the TCA cycle. Compared with the S group, red represents an increase in the NI area of the M group, and blue represents a decrease. n = 3, x ± SD. ^*P < 0.05, ^**P < 0.01. vs. S/M. S: sham surgery. M: model.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Visualization of reprogrammed FA oxidative metabolism in the heart tissue of rats with HF using AFADESI-MSI after OTCD. (A) Schematic diagram of OTCD. (B) MS images of representative FALs, FAs and oxo FAs in the heart tissues from the S and M groups. n = 3, x ± SD. ^*P < 0.05, ^**P < 0.01 vs. S/M.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Valsartan and QDP showed different metabolic regulation characteristics in heart tissue of HF rats. (A) MS images of representative metabolites in groups S, M, Q, and V, and ion intensity bar charts corresponding to different damaged microareas. (B) Violin plots of CPT2, CRAT, PDHB, and ACLY expression in the I and NI areas. n = 3, x ± SD. P < 0.05, ^**P < 0.01 vs. S/M. ^#P < 0.05, ^##P < 0.01 vs. Q/M, V/M.

  下载: 全尺寸图片 幻灯片
•