Proteoforms in Disease: Biomedical Applications of Top-Down Proteomics

Abstract

The proteome is a dynamic landscape of proteoforms arising from genetic mutations, alternative splicing, and post-translational modifications (PTMs), which collectively drive biological function and disease phenotypes. Mass spectrometry (MS)-based proteomics has emerged as an essential technique for elucidating this molecular complexity. Although bottom-up proteomics enables deep protein identification and quantification through peptide-level analysis, it disrupts molecular connectivity and introduces a peptide-to-protein inference problem, which is suboptimal for proteoform analysis. Top-down proteomics (TDP) offers a complementary approach by analyzing intact proteins, preserving molecular connectivity and enabling direct characterization and quantification of proteoforms. This capability is increasingly vital for understanding heterogeneous human diseases. Here, we review the evolving role of TDP in biomedical research, highlighting studies that revealed proteoform-level alterations, identified candidate biomarkers, and advanced our understanding of the roles of proteoforms in human diseases.

Introduction

Unlike the static nature of the genome, the proteome is composed of a dynamic network of proteoforms, a term describing the diverse protein products produced from a single gene due to genetic variations, alternative splicing, and post-translational modifications (PTMs), among others.¹ As proteins serve as the major effectors of biological processes, it is increasingly recognized that alterations at the proteoform level are closely associated with cellular stress responses and disease pathogenesis.² To gain a comprehensive understanding of the proteome, mass spectrometry (MS)-based proteomics has become essential for the study of protein abundance, structure, interactions, and PTMs.^3–8 Methodologically, the field is dominated by two primary strategies: the analysis of intact proteins in top-down proteomics (TDP) and the analysis of peptides in bottom-up proteomics (BUP).^4,9–11

As the most widely adopted and technologically mature strategy, BUP analyzes small peptides generated by the complete proteolytic digestion of proteins extracted from a biological sample.^5,6,11 The peptides are then analyzed with liquid chromatography (LC)-tandem MS (MS/MS) with high throughput and sensitivity.^5,6 The relative ease of peptide separation, ionization, and fragmentation enables the routine detection and quantification of thousands of proteins. The greater depth of coverage afforded by BUP compared to methods like 2D gel electrophoresis led to this method serving as a workhorse for proteomic analysis and, as such, this strategy has benefited from the development of robust workflows, including well-developed and broadly applicable separation methods and platforms for data analysis.⁵ Despite these advantages, complete digestion introduces a ‘peptide-to-protein inference’ problem and results in the loss of connectivity from combinatorially modified proteins with limited sequence coverage, making BUP suboptimal for proteoform characterization¹².

More recently, middle-down proteomics (MDP) has emerged as an intermediate strategy that partially preserves proteoform connectivity while retaining several advantages of peptide-based analysis.^13–16 MDP relies on restricted or site-specific proteolysis, resulting in the generation of larger peptides.^13,14 This strategy better retains short-range PTM combinations and isoform connectivity compared to BUP, while circumventing the ionization and charge-state dispersion challenges associated with the analysis of proteins. However, MDP remains constrained by challenges associated with reproducible proteolysis, dependence on protein sequence and protease specificity, stringent digestion conditions, and limited computational support. Despite the maturity of BUP or the potential of MDP, these peptide-centric approaches can still obscure key molecular details important for biological function.

In recent years, TDP has gained prominence as an increasingly important approach for proteoform characterization.^{1,2,4,7,8,17–20} In contrast to BUP and MDP, TDP bypasses proteolytic digestion to analyze proteins in their intact state, preserving the molecular connectivity across the entire molecule and enabling the characterization and quantification of proteoforms.⁴ Following the initial mass measurement, subsequent isolation and fragmentation of proteoforms of interest allows for identification and localization of sequence variations and PTMs. However, while TDP enables the direct analysis of intact proteoforms, it is challenged in key areas such as sensitivity, solubility, dynamic range, data analysis, and proteome complexity.⁸ Regardless, as these limitations are progressively addressed, the capability of TDP to resolve the intricate proteoform landscape is becoming increasingly valuable for biomedical research, particularly when diverse biological mechanisms drive a disease, and genotype alone is unable to explain phenotypic heterogeneity.^21–23 Proteoform-level analysis bridges this gap by providing detailed molecular insights into pathophysiology and phenotype.

Importantly, as TDP continues to mature, its application to biologically relevant systems has provided new insights into disease biology and accelerated the discovery of potential biomarkers. This review will provide a comprehensive perspective of the role of TDP across diverse biomedical applications and highlight studies that have advanced our understanding of proteoform-disease correlations (Figure 1).

Figure 1. Interrogating the human proteoform landscape in health and disease.

Schematic representation of TDP applications. By preserving the integrity of the protein molecule, TDP facilitates the mapping of post-translational modifications (PTMs), quantification of disease-specific proteoforms, and structural characterization. These capabilities are being leveraged to identify novel biomarkers and therapeutic targets across various organ systems and disease states. A summary of notable proteoforms can be found in Supplemental Table 1. Abbreviations: UVPD: ultraviolet photodissociation; CID: collision induced dissociation; ExD: electron-activated dissociation; m/z: mass to charge; FC: fold change.

Biomedical Applications

Cardiovascular Diseases

Cardiovascular diseases (CVDs) remain the leading cause of death globally with major contributions to disability and growing health care costs.²⁴ With the global burden of CVD projected to rise in the coming decades,²⁵ it has become increasingly important to identify the underlying mechanisms and work towards discovering more effective therapeutic options. Proteomics enables comprehensive molecular characterization of proteome changes associated with CVDs, aiding biomarker discovery and the identification of therapeutic targets.^26–28

In the early application of quantitative TDP to CVD, research largely focused on myofilament proteins, which play central roles in regulating cardiac contraction and are highly modulated by proteoform-level changes.²⁹ Cardiac troponin I (cTnI), the inhibitory subunit of the troponin complex that regulates cardiac contraction and relaxation by modulating myofilament Ca²⁺ sensitivity,^30,31 is a well-established biomarker for cardiac injury.³² Zhang et al. employed immunoaffinity purification in combination with direct-infusion TDP to characterize cTnI proteoforms in the hearts of patients with mild hypertrophy, severe hypertrophy/dilation, and end-stage heart failure (HF).³³ Such analysis detected a total of 22 cTnI proteoforms across all tissues, including a number of truncated products, as well as mono- and bis-phosphorylated proteoforms. Interestingly, quantitative analysis revealed a decrease in cTnI phosphorylation in all groups relative to nonfailing controls. Critically, the reduction in cTnI phosphorylation correlated with the severity of the phenotype, with a mild, albeit statistically significant, decrease in hearts exhibiting mild hypertrophy and the greatest decrease in end-stage HF hearts. MS/MS also confirmed the phosphorylated residues in cTnI as Ser22 and Ser23, two consensus protein kinase-A sites, and further suggested that phosphorylation of Ser23 requires prior phosphorylation at Ser22. In a separate study, Tiambeng et al. developed a novel enrichment strategy that coupled peptide-functionalized superparamagnetic nanoparticles with TDP for the unbiased characterization of cTnI directly from serum.³⁴ This high sensitivity approach depleted abundant serum proteins and enabled the detection of as little as 0.75 ng/mL cTnI, which is within the range of serum levels following acute myocardial infarction (AMI; ~0.5–50 ng/mL).^35,36 Use of this enrichment also allowed detailed characterization of cTnI proteoforms, revealing proteoform-specific molecular fingerprints associated with cardiovascular disease etiologies. Collectively, these findings provided insight into coordinated regulation of the phosphorylation sites in cTnI and also underscored the potential of PTMs as disease biomarkers.

A leading cause of HF is AMI, which irreversibly damages the heart, leading to compensatory remodeling that eventually turns maladaptive.³⁷ To better understand the proteoform changes occurring early after AMI, Peng et al. carried out TDP analysis of cardiac tissue from a swine model of AMI.³⁸ The use of an online LC-MS TDP approach with one-dimensional reverse phase LC (RPLC) allowed for the detection of multiple sarcomeric proteins, including cTnI, cardiac troponin T, α-tropomyosin, the ventricular myosin regulatory light chain in the myofilament, as well as enigma homolog 2 (ENH2) in the Z-disc. Notably, subsequent fraction collection and offline MS/MS analysis revealed a concerted reduction in the phosphorylation of cTnI (Ser22/23), ventricular myosin regulatory light chain (Ser14), and ENH2 (Ser118), suggesting coordinated crosstalk between the myofilament and Z-disc in swine hearts immediately following AMI. A finding that revealed, for the first time, crosstalk between signaling in the myofilaments and Z-disc.

Using a similar TDP strategy featuring RPLC and online LC-MS/MS analysis, Chapman et al. assessed proteoform alterations in human heart tissue from patients with HF due to ischemic cardiomyopathy (ICM).³⁹ Although this study examined end-stage human failing hearts, several proteoforms exhibited similar alterations to those detected in the swine AMI model, like the concerted reduction in the phosphorylation of cTnI and ENH2. The greater proteome depth afforded by improved instrumentation facilitated the detection of further alterations in Z-disc proteoforms, including increased phosphorylation of both calsarcin-1 and muscle LIM protein in the myocardium of patients with ICM relative to that in nonfailing controls. Interestingly, an increase in the phosphorylation of cardiac troponin T was also detected in the hearts of human ICM patients that was not detected in the swine AMI model, which may be due to differences in either species or the stage of disease (immediately following AMI in swine versus end-stage HF in humans). Enabled by the photocleavable surfactant, 4-hexylphenylazosulfonate (Azo),⁴⁰ Rogers et al. characterized phospholamban⁴¹, a key transmembrane regulator of sarcoplasmic reticulum Ca²⁺ handling whose PTM state modulates cardiac contraction and relaxation, in myocardial tissue from patients with ICM.⁴² Quantitative analysis revealed a significant reduction in phospholamban phosphorylation in human ICM. As phospholamban is a major downstream target of protein kinase A,⁴² its dephosphorylation suggests dysregulation of β-adrenergic pathway and the sarcomere-Ca²⁺-handling axis.

Hypertrophic cardiomyopathy (HCM) is a prevalent heart disease characterized by left ventricular hypertrophy, arrythmias, and sudden cardiac death.^43,44 Early studies linking pathogenic variants in genes encoding components of the sarcomere gave rise to the view of HCM as a “disease of the sarcomere.”⁴⁵ There are currently eight sarcomeric genes with strong evidence for pathogenicity in HCM, including MYH7, MYBPC3, TNNT2, TPM1, MYL3, MYL2, TNNI3, ACTC1, and TNNC1;⁴⁶ however, it is not known how disease-associated variants in these genes give rise to similar clinical phenotypes. To address this question, Tucholski et al. performed RPLC-MS-based TDP analysis of tissue from HCM patients undergoing septal myectomy surgery to relieve outflow tract obstruction and location-matched tissue from nonfailing controls.⁴⁷ This TDP analysis revealed a complex landscape of sarcomeric proteoforms, with over 45 sarcomere proteoforms detected arising from genetic variants, alternative splicing, and PTMs in donor and HCM tissues. Quantitative TDP revealed decreases in phosphorylated proteoforms of cTnI, ENH2, α-tropomyosin, and ventricular myosin regulatory light chain, while those for cardiac troponin T were increased (Figure 2). Surprisingly, the detected proteoform alterations were consistent across patient genotypes, indicating a convergent effect of HCM remodeling that is independent of the genetic variants. The observed convergence of relative proteoform-level alterations across HCM genotypes suggests the activation of common disease pathways that may be targeted therapeutically to treat this genetically heterogeneous disease.

Figure 2. Altered phosphorylation of cTnT, Tpm1.1, and MLC-2v.

Representative deconvoluted mass spectra from donor hearts (black) and HCM tissues (red) for (A) cTnT, (B) Tpm1.1, and (C) MLC-2v. Monophosphorylation is denoted with red p. Total phosphorylation (Ptot) calculated by mol Pi/mol protein for (D) cTnT, (E) Tpm1.1, and (F) MLC-2v for Ctrl (n = 16) and HCM (n = 16). Horizontal bars represent the mean of the group and error bars represent SEM (gray, Ctrl; black, HCM). Groups were considered statistically different at P < 0.05. Reproduced from Ref 47; Tucholski et al. Proc. Natl. Acad. Sci. U.S.A. 2018. 117 (40) 24691–24700.

In recent years, the view of HCM as a “disease of the sarcomere” has been challenged with growing recognition of contributions from outside the sarcomere to disease pathogenesis and heterogeneity.^48–50 To more broadly assess proteoform alterations in HCM, Gao et. al. recently employed a serial extraction method utilizing the photocleavable surfactant, Azo,⁴⁰ in combination with online RPLC-MS-based TDP.⁵¹ This strategy greatly expanded proteoform coverage with approximately 2,000 cardiac proteoforms being detected from proteins localized across multiple cellular compartments, including the sarcomere, nucleus, mitochondria, and sarcoplasmic reticulum. Quantitative comparison revealed similar changes in the sarcomeres to those detected previously, further confirming prior findings. In addition, a host of other proteoform changes were also uncovered. Beyond the anticipated dysregulation of phosphorylation, analysis also revealed marked decreases in total succinylation and acetylation of mitochondrial and nucleosome proteoforms, respectively. Notably, a subset of observed alterations was correlated to interventricular septal thickness and systolic left ventricular internal diameter, demonstrating that clinical features pertinent to cardiac disease pathophysiology can be linked to molecular-level changes with TDP.

Beyond tissue-based studies focused on late-stage CVD, Wilkins et al. used TDP to characterize apolipoproteins (ApoAI and ApoAII), abundant serum proteins involved in the transport of lipids, and identify associations between proteoforms and human cardiometabolic phenotypes.⁵² Through serum proteoform profiling of 150 individuals, TDP analysis detected 15 ApoAI and 9 ApoAII proteoforms of varying abundance, characterized by diverse modifications including acylation, oxidation, glycation, cysteinylation, and specific truncations. Paired comparison of samples collected from participants in the Coronary Artery Risk Development in Young Adults (CARDIA) cohort at years 20 and 25 revealed positive and negative correlations with indices such as high-density lipoproteins, body mass index, waist circumference, and fasting serum glucose. Such associations suggest that circulating ApoAI/ApoAII proteoforms may drive or reflect dysregulated biological pathways underlying cardiometabolic health, further emphasizing the importance of proteoform-level characterization and quantification with TDP.

Cancer

As the second leading cause of death in the United States,⁵³ cancer presents a critical challenge that TDP is uniquely positioned to address. Given the established link between proteoforms and oncogenesis,^54,55 TDP has emerged as a robust platform for biomarker discovery and pharmacological profiling. The RAS family of oncogenes, which includes KRAS, NRAS, and HRAS, is frequently mutated in human cancer and, thus, has diagnostic potential.⁵⁶ Ntai et al. developed a targeted TDP workflow incorporating immunoaffinity enrichment with nanocapillary RPLC-MS/MS for the allele- and isoform-specific interrogation of proteoforms expressed from KRAS genes within human colorectal cancer cell lines and tumor samples (Figure 3).⁵⁷ This study revealed mutation/PTM cross-talk through mapping of gene-, isoform-, and allele-specific proteoform patterns. This includes discovering the cell-line specific nitrosylation of cysteine 118 on the non-mutant proteoform upon knockout of the pathogenic KRAS G13D variant, revealing tissue-level differences in C-terminal carboxymethylation, and determining the stoichiometric ratio between proteoforms expressed from mutant and non-mutant alleles.

Figure 3. KRAS4b proteoform detection in cell lines.

(A) Isotopically resolved mass spectrum of KRAS4b, charge state 24+, isolated from three isogenic cell lines (DLD-1 PAR, DLD-1 WT, DLD-1 MUT) and illustrating the effect of KRAS mutational status on posttranslational modifications. Asterisks denote known electrospray oxidation products. (B) Schematic of KRAS4b and its PTM profile. (C) Depiction of the GDP/GTP binding pocket within the tertiary structure of KRAS4b (yellow, Cys118; magenta, Asp13; cyan, GDP; green, Mg²⁺). This figure was based on PDB ID code 4OBE. Reproduced from Ref 57; Ntai et al. Proc. Natl. Acad. Sci. U.S.A. 2018. 115 (16) 4140–4145.

In a follow-up study, Adams et al. further explored isoform-specific PTMs in KRAS4a and KRAS4b across 14 cell lines and 34 colorectal tumor samples.⁵⁸ Their analysis greatly expanded the number of curated RAS proteoforms from 11 to 39, including the identification of novel KRAS4B proteoforms lacking the C185 residue in the COR-L23 cell line and many primary colorectal tumor samples. Interestingly, further analysis revealed that these proteoforms are unable to activate the MAPK signaling pathway, a signaling cascade that has critical roles in regulating multiple cellular activities including cell proliferation and survival.⁵⁹ Although the mechanism by which this modification is generated remains to be determined, this proteoform may act as an anti-proliferative regulator or dominant-negative inhibitor, offering a promising avenue for therapeutic development.

In analyzing isogenic human nonmetastatic (SW480) and metastatic (SW620) colorectal cancer cell lines, McCool et al. identified over 23,000 proteoforms derived from over 2,300 genes.⁶⁰ Of these, many are involved in signaling pathways related to colorectal cancer, including the aforementioned MAPK pathway, as well as the WNT/β-catenin pathway (governing tissue development and homeostasis),⁶¹ the Akt pathway (regulating survival and growth),⁶² and the mTOR pathway (controlling cellular growth and metabolism).⁶³ To achieve this extensive coverage, they employed a suite of capillary zone electrophoresis (CZE)-MS approaches, ranging from 1D CZE-MS/MS to 2D and 3D configurations incorporating size exclusion chromatography (SEC) and RPLC. In particular, upon comparison of the nonmetastatic and metastatic cell lines, the proteoform profiles diverged, with 460 proteoforms having significantly different abundances. Further analysis revealed that different proteoforms from the same genes, including some with phosphorylation, were regulated differently and had opposite expression patterns between the two cell lines. This includes proteoforms derived from CALM1, HDGF, JPT1, NPM1 and DAP, which have varying contributions to metastasis, transcriptional regulation, and the modulation of cell death. In a similar study, Fang et al. focused specifically on differences in histone proteoforms in the SW480 and SW620 cell lines, as the development of colorectal cancer is associated with epigenetic alteration and regulation. Histones were extracted from isolated nuclei, separated by SDS-polyacrylamide gel electrophoresis, gel extracted, and subject to further separation and analysis using CZE-MS/MS. Using this strategy, a total of 230 histone proteoforms were detected, among which 34 were significantly altered in abundance in the SW620 cell line.⁶⁴ Together, these findings identified specific proteoforms implicated in the modulation of metastasis that could potentially be used as biomarkers for colorectal cancer.

Protein kinase inhibitors are a common class of anti-cancer agents, yet there is a lack of insight on how such molecules interact with or otherwise impact targeted proteins and the wider proteome.⁶⁵ D’Ippolito et al. developed an in vitro TDP assay to evaluate how the KRAS4B protein is modified in response to treatment with compounds from the UCSF Small Molecule Discovery Center library and FDA-approved kinase inhibitors.⁶⁶ Their TDP RPLC-MS/MS approach enabled the evaluation of binding stoichiometry, specificity, localization, and characterization of reaction products. Importantly, this method is adaptable and can accelerate the development and evaluation of KRAS4B-targeted inhibitors. In looking at the effects of a broad-spectrum kinase inhibitor, staurosporin, Chowdhury et al. quantified nearly 1,200 proteoforms from HeLa cells using RPLC-MS/MS.⁶⁷ In comparison to a DMSO-treated control, there were 55 proteoforms significantly changed following staurosporin treatment, many of which were related to metabolic and apoptotic signaling pathways. Holt et al. further demonstrated the utility of TDP in characterizing combinatorially modified proteins by measuring histone proteoform alterations in triple-negative breast cancer cells.⁶⁸ Using a reproducible method combining nuclei isolation, offline LC-based histone purification and TDP RPLC-MS/MS, they tracked multiple PTMs following treatment with butyrate, a deacetylase inhibitor. The development of such assays highlights the critical role of TDP in providing proteoform-level validation of target engagement; by preserving the connectivity of multiple PTMs, TDP can offer a comprehensive assessment that traditional BUP approaches cannot resolve.

Extending beyond the denaturing LC-MS approaches previously described, native top-down mass spectrometry (nTDMS), combining TDP with native MS, has emerged as a powerful technique to probe higher order structures and assembly. Gomes et al. developed a nTDMS approach, comprised of offline native SEC, nano-electrospray ionization, and field asymmetric ion mobility spectrometry (FAIMS), to identify protein complexes formed by monomeric proteoform arrangements of one or more gene products (complexoforms).⁶⁹ In estrogen receptor-alpha (ERα)-positive MCF-7 breast cancer cells and endocrine-resistant breast cancer cells (overexpression of EGFR), they identified and characterized 104 complexoforms, which consisted of proteoform-metal, proteoform-proteoform–metal, and proteoform-proteoform assemblies. Notably, in the endocrine-resistant cell line, nuclear transport factor 2 (NUTF2) was found to have reduced dimerization, disrupting its role as a shuttle protein for nuclear import. While NUTF2 overexpression diminished cancer cell growth, they showed that mutations at specific PTM sites bidirectionally modulated this effect, highlighting the necessity of proteoform-level analysis to resolve these regulatory switches. As nTDMS continues to be employed in biomedical research, crucial insights into the structure and dynamics of endogenous protein complexes will be uncovered, further elucidating their role in disease pathogenesis.

To drive spatially-resolved proteoform characterization, Su et al. developed proteoform imaging MS (PiMS), a technique that couples individual ion MS with nanospray desorption electrospray ionization to analyze intact proteins directly from tissue microenvironments with <80-μm spatial resolution.⁷⁰ Building on this foundation, McGee et al. developed AutoPiMS, a semiautomated, single-ion MS-based multiplexed workflow for TDP.⁷¹ By integrating a computation engine that automates proteoform target selection and MS/MS acquisition method generation, AutoPiMS enables intact mass measurement, proteoform identification, spatial distribution, and quantitative differential expression analysis. Applying this approach to tumor and stromal regions from matched ovarian cancer biopsies from the same patient revealed over 1,000 proteoforms, 303 of which exhibited significant differential abundance between the two tissue types. A key demonstration of this method’s utility involved mapping the spatial distribution of cysteine-rich intestinal protein 1 (CRIP1), where imaging revealed distinct localization patterns: unmethylated CRIP1 enriched in stroma, whereas mono- and dimethylated forms (Arg68) localized predominantly to tumor regions.⁷¹ It is easy to envision that such an approach could be useful for defining tumor boundaries during surgical resection with further development of tumor type-specific proteoform profiles enabling differentiation of tumor cells from normal tissue.

Neurodegenerative Diseases

Neurodegenerative diseases (NDDs), including Alzheimer’s disease (AD), Parkinson’s disease (PD), and amyotrophic lateral sclerosis (ALS), among others, represent a major cause of mortality and disability in the elderly and are defined by the abnormal aggregation of proteins.⁷² The dysregulation of PTMs has been associated with the modulation of protein aggregation,⁷² making TDP an optimal technique to further elucidate pathogenesis and progression of NDDs. An early study by Kellie et al. characterized α-synuclein—the protein constituent of Lewy bodies, which are protein aggregates formed in the neurons of patients with PD—proteoforms in the frontal gyrus of patients with PD.⁷³ This TDP approach employed antibody-based enrichment of α-synuclein from SDS-soluble and insoluble protein extracts, allowing the detection of ten truncated α-synuclein proteoforms along with full-length α-synuclein in SDS-soluble fractions. Analysis of the SDS-insoluble fraction, which contains the insoluble Lewy bodies, identified five truncated proteoforms, as well as full-length phosphorylated and non-phosphorylated α-synuclein. Although several of the detected truncated proteoforms had been previously reported in the literature, seven were novel proteoforms. It is also noteworthy that, while analysis following Glu-C digestion was able to detect five of the identified α-synuclein proteoforms, six were missed using this method. The ability to characterize truncations is critically important to understanding disease, as these proteoforms can impact biological activity, subcellular localization, and function.⁷⁴ Peptide-based approaches often lose information on protein termini, owing to incomplete sequence coverage, which can result in missing terminal peptides, and the loss of connectivity between the N- and C-termini, highlighting the an important advantage of intact protein analysis for proteoform-level characterization.

Beyond truncation and phosphorylation, addition of O-linked N-acetyl-glucosamine has emerged as a potential PTM on α-synuclein that can prevent detrimental aggregation of the protein.^75,76 Recently, Miller et al. developed a combined TDP and MDP methodology coupling high-resolution trapped ion mobility spectrometry (TIMS) with electron capture dissociation and ultraviolet photodissociation, capable of site-specific characterization of α-synuclein as well as identification of O-linked N-acetyl-glucosamine positional isomers.⁷⁷ TDP unambiguously confirmed the sequence and PTM localization for each of the glycoproteoforms, supporting the presence of O-linked N-acetyl-glucosamine at Thr72, Thr75, Thr81 and Ser87 with high sequence coverage. The ability of this combined TDP and MDP approach to define α-synuclein-modified glycoproteoforms, including positional isomers, represents a crucial first step towards evaluating how these modifications affect the activity and pathological aggregation of the protein.

One of the more prominent challenges in the diagnosis and prognosis of NDDs is the lack of reliable biomarkers from non-invasive sources. Contini et al. employed TDP to screen for peripheral biomarkers of PD and AD in saliva, as it is non-invasive, has a more favorable dynamic range than plasma and contains markers that may reflect local and systemic conditions.^78,79 This study reported the identification and characterization of 51 proteoforms. This included proteoforms generated by phosphorylation, proteolysis, N-terminal acetylation, methionine oxidation, and cysteine oxidation (formation of disulfide bridges, glutathionylation, cysteinylation, and nitrosylation), in soluble acidic extracts of patient saliva. Comparison of proteoforms detected in the saliva of PD, AD, and healthy patients revealed unique proteoform signatures, demonstrating that salivary proteins harbor disease-specific PTM patterns that are detectable via TDP. Interestingly, certain proteoforms correlated with clinical parameters such as the Unified PD Rating Scale III, olfactory function, and the Montreal Cognitive Assessment, highlighting their utility as putative biomarkers to detect PD and enable its differentiation from AD.

The association of proteoforms with clinicopathological traits of AD was also explored in a global quantitative TDP study by Fulcher et al., in which they characterized proteoforms in dorsal lateral prefrontal cortex biopsies from over 100 subjects across the Religious Orders Study and Rush Memory and Aging Project (Figure 4).⁸⁰ Using a RPLC-MS/MS discovery approach that combined a differential solubilization technique with hexafluoroisopropanol, FAIMS, and a TDP quantitative software (TopPICR), they identified 11,782 proteoforms from 1,213 proteins. The incorporation of FAIMS was previously shown to improve TDP coverage of brain tissue and enabled the identification and characterization of key proteoforms implicated in NDDs, including α-synuclein, Parkinsonism-associated deglycase, splice variants of tau, and intact amyloid beta.⁸¹ Of the almost 12,000 unique proteoforms detected in the optimized approach, 154 proteoforms observed associations with at least one of the multiple clinicopathological variables tested against. Interestingly, the heterogeneity of amyloid beta proteoforms resulted in distinct differential pathological associations, which could have significant implications on the efficacy of specific therapies. In an analogous study, Askarani et al. characterized the temporal cortex of AD patients against healthy controls.⁸² In this approach, which combined SEC fractionation, CZE, and quantitative TDP, almost 3,200 total proteoforms were identified, with many appearing exclusively in either healthy or diseased samples. Of the 736 proteoforms that were quantified in both cohorts, 157 proteoforms exhibited a significant difference in abundance, identifying condition-specific proteoforms as potential biomarkers and revealing bidirectional proteoform regulation in multiple key genes that are associated with AD development through processes such as telomere organization, microtubule disruption, amyloid formation, and axogenesis.

Figure 4. Workflow for collecting and analyzing top-down proteomics datasets of 103 human brains.

Human brain tissue was processed by differential solubilization, separated into two fractions, and then analyzed by LC-MS with gas-phase fractionation using FAIMS. After database searching with TopPIC, proteoform spectrum match (PrSM) result files are imported into R for downstream analysis with TopPICR. Figure was partially created with BioRender. Reproduced from Ref 80; Fulcher et al. Mol Cell Proteomics. 2025. 24 (6) 100983.

Using nTDMS, Nshanian et al. characterized the interaction between tau, a protein that forms neurofibrillary tangles in AD, and CLR01, a synthetic compound designed to prevent abnormal protein aggregation.⁸³ In addition to determining binding stoichiometry across different tau proteoforms, the researchers employed electron capture dissociation to localize inhibitor binding sites and ion mobility for additional structural elucidation, providing potential insight into the mechanism of inhibition, which may involve structural rearrangement and compaction of the protein upon CLR01 binding. Such information would be challenging to obtain using enzymatic digestion-based strategies as retention of the ligand would likely be lost and with it, information pertaining to drug-proteoform interactions. Beyond research on AD and PD, there have been additional studies using TDP across other NDDs and broader neurophysiological contexts. Ye et al. coupled matrix-assisted laser desorption/ionization (MALDI)-MS imaging and TDP to help identify potential biomarkers for neuropsychiatric syndromes resulting from neurotransmission imbalance in early development.⁸⁴ Multiple TDP methods have also been developed to investigate disease-relevant modifications in ALS, a rapid onset motor neuron disease.⁸⁵ Specifically, Cu/Zn-superoxide dismutase (SOD1), a protein that has multiple genetic variants linked to familial ALS, has been a primary target in research investigating the stoichiometry of cross-linker binding and the impact of PTMs on protein conformation and pathogenecity.^86–88

Infectious Diseases

Infectious diseases caused by bacteria, viruses, and parasites represent clinical challenges due to antimicrobial resistance and rapid transmission.⁸⁹ In hospitals, effective management relies on the rapid identification of the pathogen, particularly drug-resistant strains. For drug or vaccine development it is also essential to have a better understanding of the underlying mechanisms and of the host immune response. TDP can be highly effective for the characterization of proteoforms that govern microbial virulence, immune evasion, and therapeutic responses. MALDI time-of-flight (TOF) MS, which is based on direct profiling of intact proteins, has become the cornerstone of clinical microbial identification in hospitals worldwide, representing the most widely adopted approach in routine diagnostics.⁹⁰ When enhanced with post-source decay (PSD) fragmentation, this platform can identify strain-specific proteoforms such as distinct Shiga toxin variants and osmotically inducible protein forms (OSmY) in pathogenic E. coli, producing cleaner spectra through low-energy ion dissociation.^91,92 However, LC-MS/MS using high-resolution instruments, such as orbitraps, offers superior discriminatory power, distinguishing bacterial strains previously considered identical by MALDI-TOF MS. This approach has identified 220 proteins and over 500 proteoforms from different E. coli strains,⁹³ characterized M. tuberculosis complex species at the proteoform level,⁹⁴ and enabled trace-level quantification of Staphylococcal enterotoxin variants at 2.5 ng/mL through parallel reaction monitoring.⁹⁵ Liquid extraction surface analysis mass spectrometry (LESA-MS) complements these methods by enabling direct analysis of live bacterial cultures with minimal preparation, identifying 39 intact proteins from mixed species within minutes⁹⁶ and distinguishing antibiotic-resistant bacterial strains directly from infected 3D skin wounds,⁹⁷ including rapid identification of ESKAPE pathogens responsible for hospital-acquired infections.⁹⁸

TDP approaches can also reveal critical biological mechanisms through comprehensive proteoform characterization. Analysis of M. tuberculosis by Parra et al. identified 130 proteoforms revealing functionally distinct lipidation states of the lipoglycoprotein LpqH, a virulence-associated protein, providing insight into infectious properties.⁹⁹ Studies of malaria liver stages identified proteoform signatures suitable for immunization candidates,¹⁰⁰ while analysis of T. vaginalis revealed 25 immunoreactive proteoforms, highlighting potential virulence factors and parasite morphogenesis.¹⁰¹ Ansong et al. characterized foodborne S. typhimurium, mapping 1,665 proteoforms from 563 unique proteins, revealing S-thiolated proteoforms that function as a unique regulatory switch under infection-like conditions.¹⁰² Proteoforms carrying specific PTMs have been shown to play crucial roles in infection mechanisms, as exemplified by the comprehensive TDP analysis of the N. meningitidis major pilin PiIE. This primary component of Type IV pili carries critical virulence-associated modifications, including extensive glycosylation^103,104 and glycerophosphorylation.¹⁰⁵ While glycosylation promotes immune evasion, glycerophosphorylation enables bacterial translocation across epithelial barriers—a pivotal step in this pathogen's virulence.^103–106

Flagella, the whip-like appendages present in many enteric and gastric pathogens, enable bacterial penetration of mucus layers and host tissues through rotational motility. Analysis across C. jejuni, H. pylori, A. caviae, and L. monocytogenes demonstrated that strain-specific O-linked glycosylation of flagellar filaments modulates not only structural integrity and motility but also antigenic variation critical for immune evasion.¹⁰⁷ Similarly, TDP distinguished gastric from enterohepatic Helicobacter species through the differential detection of nickel-binding histidine-rich proteins Hpn and Hpn-2, key factors for gastric colonization.¹⁰⁸

Viral pathogens exhibit equally complex proteoform landscapes critical for immune recognition, as demonstrated for SARS-CoV-2. Hydrophilic interaction chromatography (HILIC) coupled with MS characterized spike protein receptor-binding domain (RBD) glycoproteoforms, revealing unexpected cysteinylation and N-terminal sequence variations while capturing low-abundance, heavily glycosylated species.¹⁰⁹ Roberts et al. developed a hybrid denatured and nTDMS strategy combining FT-ICR MS/MS with TIMS to resolve structural heterogeneity of the SARS-CoV-2 spike RBD (Figure 5).¹¹⁰ This approach identified eight distinct O-glycoproteoforms and provided site-specific insight into glycan occupancy and structural variation. In a follow-up study, Roberts et al. applied TDP to the Omicron RBD and revealed variant-specific differences in core glycan composition and O-glycoproteoform utilization compared with SARS-CoV-2 strains.¹¹¹ Together, these studies demonstrate how TDP can resolve glycoproteoforms and structural heterogeneity that are difficult to capture with peptide-centric approaches, providing molecular insight relevant to viral biology, immune recognition, and therapeutic or vaccine design.

Figure 5. S protein regional-binding domain (S-RBD) O-glycoproteoform analysis by high-resolution top-down MS.

(A) MS¹ analysis of S-RBD O-glycoproteoforms following denaturation. (B) Magnified MS¹ from A showing isotopically and baseline resolved S-RBD proteoforms allowing for accurate determination of individual intact glycoproteoforms. Most abundant charge state (15+) is isolated. Theoretical isotope distributions (red circles) are overlaid on the experimentally obtained mass spectrum to illustrate the high mass accuracy. All individual ion assignments are within 1 ppm from the theoretical mass. (C) Illustration of the isolation of a specific S-RBD glycoproteoform using ultrahigh-resolution FTICR to enable glycan site and structure characterization from intact glycoproteoforms. (D) MS/MS characterization of S-RBD proteoform isolated from the quadrupole window centered at 1760.5 m/z. Assignments of the glycan structures are marked in the spectrum with the legend shown on the right side. Glycoproteoform characterization reveals the specific S-RBD proteoform to have a Core 2 type GalNAcGal(GalNeuAc)(GlcNAcGalFuc) glycan. Neutral loss glycan products are labeled. N-Terminal acetylation (Ac) is labeled and corresponds to a +42 Da mass shift. Solid star represents the 15+ charge state precursor ion corresponding to 1760.5 m/z, and hollow star represents the 15+ charge state precursor ion corresponding to 1757.7 m/z (−Ac). Asterisk “*” denotes an oxonium ion loss. Figure and associated legend reprinted and adapted with permission from Ref. 110; Roberts et al. J. Am. Chem. Soc. 2021, 143, 31, 12014–12024. Copyright 2021 American Chemical Society.

Beyond characterizing pathogens, TDP can also capture host immune responses^112,113 and support therapeutic drug monitoring. Analysis of intact IgG1 antibodies revealed that human plasma IgG1 repertoires constitute surprisingly simple yet individually unique proteoform signatures that dynamically respond to infectious challenge.¹¹³ Unlike traditional serology that measures bulk antibody titers, this approach captures the complete molecular diversity of circulating antibodies, including subclass distributions, glycosylation variants, and chain pairings. In patients with sepsis, disease-associated immunoglobulin proteoforms that correlate with clinical outcomes were found, demonstrating that systemic infection manifests as distinct proteoform profiles. This capability extends to specific pathogen identification through “next-generation serology,” wherein analysis of COVID-19 convalescent plasma enabled precise readout of the SARS-CoV-2 antibody repertoire.¹¹⁴ By identifying patient-specific IgG1 proteoforms induced by viral infection, infected individuals were distinguished from naive ones, demonstrating that the evolution of humoral immunity can be captured at the molecular level. Recent characterization of the “dark serum proteome” by Bondt et al. specifically resolved personalized features of IgG1 and IgA1 repertoires in severe COVID-19 patients, establishing patient-specific proteoform signatures with potential to stratify disease severity and outcomes.¹¹⁴ Likewise, TDP characterization of the C. albicans immunome in candidemia patients has identified specific proteoforms of fungal antigens differentially recognized by patient antibodies, extending these serological applications to fungal pathogens.¹¹⁵

Native MS also extends this serum proteoform profiling beyond antibodies to the broader circulating proteome. Čaval et al. revealed that individual donor plasma α-1-antichymotrypsin exhibits unique glycoproteoform landscapes that undergo extensive remodeling following septic episodes, demonstrating that acute infection induces systemic proteoform-level changes detectable in patient serum.¹¹⁶ High-resolution nTDMS further charts this individualized proteoform landscape across serum proteins, establishing that each donor possesses a distinctive molecular signature of circulating proteoforms that can be monitored as a personalized biomarker of health and disease states.¹¹⁷ Complementing this host perspective, nTDMS of SARS-CoV-2 nucleoprotein revealed that differential phosphorylation generates distinct proteoforms (phosphorylated nucleoproteins, pNPs) with specialized interaction capabilities.¹¹⁸ These studies demonstrate that nTDMS captures infection-induced proteoform remodeling across both antibody and non-antibody serum components, providing a comprehensive window into host-pathogen interactions at the molecular level.

TDP and native analyses of intact proteins demonstrate that proteoform heterogeneity actually constitutes the functional machinery driving infection outcomes. These approaches enable definitive discrimination of closely related pathogens, mapping functionally important PTMs, and detailed characterization of individual immune responses, providing essential complements to classical methods for comprehensive infectious disease management. As clinical MS platforms mature, integrating these proteoform-resolved analyses into routine infectious disease workflows promises to transform pathogen identification from a taxonomic exercise into a functional, patient-specific molecular diagnosis.

Autoimmune Diseases

Autoimmune diseases comprise a clinically diverse group of conditions that affect individuals of all ages. Driven by aberrant B-cell and T-cell activation, severe immune dysregulation causes the body to attack itself.¹¹⁹ Among these diseases, type 1 diabetes (T1D), characterized by the destruction of insulin-producing β-cells in the pancreatic islets, affects over 8 million people globally.¹²⁰ Using human islets treated with proinflammatory cytokines to model T1D-onset, Ives et al. applied an ion-mobility enabled RPLC-MS TDP to evaluate early proteoform alterations.¹²¹ Their approach, which incorporates FAIMS, resulted in the identification of over 1,600 proteoforms, 900 of which were quantifiable across control and cytokine treated samples. In addition to providing a baseline proteoform landscape of human islets, they observed 85 proteoforms that were differentially expressed (46 downregulated, 39 upregulated) post-cytokine treatment, which included proteoforms of insulin and glucagon, providing candidate biomarkers for T1D progression or therapeutic intervention. Moreover, as in the studies discussed above, TDP enabled the characterization of numerous truncated proteoforms, a form of proteolytic processing with major implications for hormone activity and function.

Multiple sclerosis is a chronic autoimmune disease characterized by degradation of the protective myelin sheath that covers nerves in the brain and spinal cord.¹²² Prompted by the lack of a single diagnostic test for multiple sclerosis, Manconi et al. investigated the usefulness of salivary proteoform signatures for differentiating individuals with multiple sclerosis from healthy controls.¹²³ Following the quantitative analysis of 119 salivary peptides and proteins, 27 exhibited a significant difference between healthy and diseased samples, with notable representation among cystatins, a family of proteins involved in both inflammation and immune response. Following a similar rationale, Olianas et al. later performed the first TDP analysis of saliva from patients with autoimmune hepatitis (AIH) and primary biliary cholangitis (PBC), two autoimmune liver diseases that result in inflammation, fibrosis, and necrosis, which ultimately lead to cirrhosis and liver failure.¹²⁴ Immunopathological differences were detected in saliva across AIH, PBC, and control groups, yielding candidate biomarkers within multiple protein families (histatin, cystatin, S100, etc.) that are capable of discriminating the two diseases from controls or also each other. Ultimately, these studies highlight the power of TDP to identify non-invasive, peripheral proteoform signatures capable of improving diagnosis, patient stratification, and disease monitoring across autoimmune disorders.

Additional Diseases and Other Biomedical Applications

TDP has been instrumental in characterizing proteoform diversity across a wide array of other pathological diseases and biomedical applications. Forte et al. sought to define the proteoform signature of patients at various stages of liver cirrhosis, a disease affecting up to 5 million in the United States.¹²⁵ Their primary objective was to identify potential diagnostic markers for risk stratification, specifically to pinpoint patients at risk of decompensation and to assist with intervention or mitigation of disease progression. To achieve this, they utilized PEPPI-MS¹²⁶ (polyacrylamide-gel-based prefractionation for analysis of intact proteoforms and protein complexes by mass spectrometry) to analyze plasma samples from a cohort of 30 patients. The study compared three distinct clinical stages: compensated, compensated with portal hypertension, and decompensated cirrhosis to map molecular changes across the spectrum of the disease. This approach led to the detection of almost 2,900 proteoforms from 99 proteins. Excluding hemoglobin subunits, 209 of these proteoforms were differentially expressed with 167 proteoforms discriminating the different stages of cirrhosis (of which 115 are enriched in the liver) (Figure 6). Although there are many proteoforms of interest, researchers highlighted the most abundant proteoforms, including ApoAI (truncation, methionine oxidation, and alpha-amino acetylation), fibrinogen (truncation), and haptoglobin (truncation and alpha-amino acetylation), which may serve as biomarkers to track disease progression. TDP analysis of apolipoproteins have also been examined in other biomedical research, including the previously described cardiometabolic phenotype study,⁵² analysis of human serum/plasma,^127–129 and association to type 2 diabetes.^130,131

Figure 6. Display of proteoforms stratified by expression in each stage of cirrhosis.

(A) differentially expressed proteoforms (DEPs) in stages II and III disease with stage I disease (compensated) used as the reference. Proteoforms upregulated in stage II and III (red square), upregulated only in stage III (orange square), downregulated in stages II and III (blue square), downregulated in stage III only (light purple square), downregulated in stage III but upregulated in stage II (dark purple square), upregulated only in stage II (green square), and downregulated only in stage II (yellow square). Proteoforms deriving from proteins enriched in liver are highlighted and grouped by color according to the figure legend. (B) profiles of proteoforms upregulated in stages I and II and downregulated in stage III (dark purple), upregulated in I and downregulated in III (light purple), upregulated in I and downregulated in II and II (light blue), downregulated in I and upregulated in II (green), upregulated in I and downregulated in II (yellow), downregulated in I and upregulated in II and III (red), and downregulated in I and upregulated in III (orange). The proteins from which DEPs derive and the relative number of proteoforms (in parenthesis) are shown. Proteins with the highest number of DEPs for each group are shown in bold. Reproduced from Ref 125; Forte et al. Mol Cell Proteomics. 2024. 23 (12) 100876.

To improve donor kidney utilization and identify molecular signatures that distinguish organ quality, Ctortecka et al. recently applied PiMS to characterize tissue from living donors, deceased donors (following brain or cardiac death), and discarded human kidneys.¹³² This study identified four key proteoforms (ACTG1, acetylated CRYAB, PARK7, and S100A4) of the 784 detected that were elevated in the marginal organs compared to living donors, providing a molecular framework for assessing organ quality. Notably, these proteoforms reflect critical pathways of oxidative stress, cellular injury, and immune activation and can serve as an objective way to stratify viable organs for transplant. In tandem, researchers also used LC-MS/MS to examine thousands of proteoforms from peripheral blood mononuclear cells collected prior to kidney transplant and 7 days post-transplant. This comparison demonstrated that donor kidney quality drives the recipient’s immune response during the first week post-transplant; although recipients share similar baseline profiles, by day 7 their molecular signatures diverge into distinct immunological trajectories.

Future Outlook and Challenges

As MS and separation technologies continue to advance, TDP is poised to become an increasingly valuable technique for interrogating disease biology. By preserving intact proteoforms, TDP provides a “bird’s-eye view” of the proteoform landscape and enables the characterization of complex systems, offering a level of molecular detail that BUP and other techniques cannot match. Despite its immense potential, several challenges remain before TDP can be widely implemented in the study of disease.^4,8 These challenges can be broadly categorized into two areas: expanding proteoform coverage and advancing the characterization and quantification of biologically and clinically relevant proteoforms.

Although observing a greater number of proteoforms does not directly translate into biological or clinical significance, expanding the detection capabilities of TDP maximizes the opportunity to uncover biologically relevant targets. Compared with BUP, the number of identified proteins in TDP remains substantially lower. This limitation is particularly apparent for proteoforms larger than 30 kDa. As protein molecular weight increases, the corresponding decrease in signal-to-noise¹³³ complicates both detection and effective sequence mapping by MS/MS fragmentation. Furthermore, although increasingly robust data analysis tools have improved high-confidence identification,^134–137 substantial advances in TDP software development are still needed. Additionally, persistent challenges such as the dynamic range of complex biological samples and the overall sensitivity required to detect low-abundance biomarkers continue to define the current frontier of TDP. Improvements in front-end separations, ion activation methods, high-sensitivity instrumentation, and computational tools will be essential for expanding TDP coverage, characterization and throughput.

While expanding proteoform coverage remains at the forefront of TDP advancement, the reliable quantification of biologically and clinically relevant proteoforms is also a critical consideration for clinical translation.¹³⁸ Currently, TDP relies heavily on label-free quantification using MS1-level intensities or integrated peak areas derived from extracted ion chromatograms. Although these methods provide a robust means for determining relative proteoform abundances among different biological conditions in discovery-phase studies, the translation of proteoform-based biomarkers into clinical diagnostic tests will require highly accurate absolute quantification and rigorous analytical frameworks.^139–141 Achieving this goal will require well-characterized proteoform reference standards capable of recapitulating complex combinations of PTMs, an area that remains substantially underdeveloped within the field of TDP.^138,142 Consequently, comprehensive characterization and accurate quantification of disease-associated proteoforms will continue to require a multi-strategy approach.¹⁴³ Integration with complementary approaches, including BUP/MDP, ion mobility, native MS, imaging MS, and targeted proteoform assays, will further enhance the ability to connect proteoform structure, abundance, localization, and function. As quantitative, analytical, and computational strategies continue to mature, TDP is expected to play an increasingly important role in defining disease-associated molecular signatures and bridging the gap between the genome and patient phenotypes.