Serum biomarkers at disease onset for personalized therapy in multiple sclerosis.

The potential of combining serum neurofilament light chain (sNfL) and glial fibrillary acidic protein (sGFAP) levels to predict disability worsening in multiple sclerosis (MS) remains underexplored. We aimed to investigate whether sNfL and sGFAP values identify distinct subgroups of patients according to the risk of disability worsening and their response to disease-modifying treatments (DMTs). This multicentre study, conducted across thirteen European hospitals, spanned from July 15, 1994, to August 18, 2022, with follow-up until September 26, 2023. We enrolled MS patients who had serum samples collected within 12 months from disease onset and before initiating DMTs. Multivariable regression models were used to estimate the risk of relapse-associated worsening (RAW), progression independent of relapse activity (PIRA), and Expanded Disability Status Scale (EDSS) score of 3. Of the 725 patients included, median age was 34.2 years (IQR, 27.6-42.4), and 509 patients (70.2%) were female. Median follow-up duration was 6.43 years (IQR, 4.65-9.81). Higher sNfL values associated with an elevated risk of RAW (HR of 1.45; 95% CI 1.19-1.76; P < 0.001), PIRA (HR of 1.43; 95% CI 1.13-1.81; P = 0.003), and reaching an EDSS of 3 (HR of 1.55; 95% CI 1.29-1.85; P < 0.001). Moreover, higher sGFAP levels were linked to a higher risk of achieving an EDSS score of 3 (HR of 1.36; 95% CI 1.06-1.74; P = 0.02) and, in patients with low sNfL values, to PIRA (HR of 1.86; 95% CI 1.01-3.45; P = 0.04). We further examined the combined effect of sNfL and sGFAP levels. Patients with low sNfL and sGFAP values (NLGL) exhibited a low risk of all outcomes and served as reference. Untreated patients with high sNfL levels showed a higher risk of RAW, PIRA, and reaching an EDSS of 3. Injectable or oral DMTs reduced the risk of RAW in these patients but failed to mitigate the risk of PIRA and reaching an EDSS of 3. Conversely, high-efficacy DMTs counteracted the heightened risk of these outcomes, except for the risk of PIRA in patients with high sNfL and sGFAP levels. Patients with low sNfL and high sGFAP values (NLGH) showed an increased risk of PIRA and achieving an EDSS of 3, which remained unchanged with either high-efficacy or other DMTs. In conclusion, evaluating sNfL and sGFAP levels at disease onset in MS may identify distinct phenotypes associated with diverse immunological pathways of disability acquisition and therapeutic response.

Progression independent of relapse activity can be predicted by passively acquired tapping speed through a smartphone for 1 month: A prospective study.

BACKGROUND: Tapping speed (TS) correlates with baseline disability scales in people with multiple sclerosis (pwMS). OBJECTIVE: The study aimed to address if progression independent of relapse activity (PIRA) could be predicted by first-month measurement of TS. METHODS: Prospective study including pwMS in one referral MS center. Consecutive patients were included and keys/second (Keys/s) were passively measured each day using an in-house smartphone application for 1 month. Median, mean, and maximum keys/s were obtained. Multivariate logistic regression models (including keys/s, age, sex, and baseline disability scores) were obtained for prediction of a PIRA event after 1 year. RESULTS: Overall, 59 patients were included in the final analysis (64.4% women, median age of 44.5 years). However, 10 patients presented a PIRA event, without differences regarding baseline characteristics between PIRA and no-PIRA groups. PIRA group presented lower median keys/s (2 vs 4 keys/s, p = 0.002) and mean keys/s (2.8 vs 4.6, p = 0.008), while maximum keys/s were similar (p = 0.32). A median ⩽ 3 keys/s was independently associated with PIRA (aOR = 16.8, p = 0.03), as did a mean ⩽ 3.7 keys/s (aOR = 17, p = 0.02). These differences were not detected regarding other variables analyzed. CONCLUSION: Low median or mean keys/s obtained during initial month of assessment were indicative of a PIRA event within the next year.

Establishing Normal Serum Values of Neurofilament Light Chains and Glial Fibrillary Acidic Protein Considering the Effects of Age and Other Demographic Factors in Healthy Adults.

Multiple studies have shown the importance of blood-based biomarkers indicating axonal damage (serum neurofilament light chains [sNfL]) or astroglia activation (serum glial fibrillary acidic protein [sGFAP]) for monitoring different neurological diseases. However, normal values of these variables remain to be clearly defined, partly due to the influence of different demographic factors. We investigated demographic differences in a cohort of healthy volunteers. A cross-sectional study was conducted including 116 healthy controls with ages between 18 and 69 years (67.5% females; n = 79). sNfL and sGFAP concentrations were measured using single-molecule arrays. Age and body mass index affected sNfL values, and age was found to be the most important factor. The normal values changed with age, and we established normal values for individuals younger than 45 years as <10 pg/mL and for controls older than 45 years as <15 pg/mL. We established normal values at <10 pg/mL for individuals younger than 45 years and <15 pg/mL for older individuals. Alternatively, a Z-score of 1.5 was relevant for all controls. sGFAP was only affected by age. Differences in normal values were evident by 55 years. The highest normality limit for sGFAP was 140 pg/mL for controls under 55 years and 280 for older controls. We defined normal levels for sNfL and sGFAP and their corresponding age-associated changes. These data may contribute to the application of such variables in clinical practice.

Different phenotypic outcome due to site-specific phosphorylation in the cancer-associated NQO1 enzyme studied by phosphomimetic mutations.

Protein phosphorylation is a common phenomenon in human flavoproteins although the functional consequences of this site-specific modification are largely unknown. Here, we evaluated the effects of site-specific phosphorylation (using phosphomimetic mutations at sites S40, S82 and T128) on multiple functional aspects as well as in the structural stability of the antioxidant and disease-associated human flavoprotein NQO1 using biophysical and biochemical methods. In vitro biophysical studies revealed effects of phosphorylation at different sites such as decreased binding affinity for FAD and structural stability of its binding site (S82), conformational stability (S40 and S82) and reduced catalytic efficiency and functional cooperativity (T128). Local stability measurements by H/D exchange in different ligation states provided structural insight into these effects. Transfection of eukaryotic cells showed that phosphorylation at sites S40 and S82 may reduce steady-levels of NQO1 protein by enhanced proteasome-induced degradation. We show that site-specific phosphorylation of human NQO1 may cause pleiotropic and counterintuitive effects on this multifunctional protein with potential implications for its relationships with human disease. Our approach allows to establish relationships between site-specific phosphorylation, functional and structural stability effects in vitro and inside cells paving the way for more detailed analyses of phosphorylation at the flavoproteome scale.

Non-severe immunosuppression might be associated with a lower risk of moderate-severe acute respiratory distress syndrome in COVID-19: A pilot study.

The role of immunosuppression among coronavirus disease 2019 (COVID-19) patients has not been elucidated and management may be challenging. This observational study included confirmed COVID-19 patients. The primary endpoint was the development of moderate-severe acute respiratory distress syndrome (ARDS). Time to moderate-severe ARDS, the need for mechanical or noninvasive ventilation (MV/NIV), death, and a composite of death or MV/NIV were secondary endpoints. Of 138 patients included, 27 (19.6%) were immunosuppressed (IS) and 95 (68.8%) were male, with a median (IQR) age of 68 (54-78) years. A significantly lower proportion of IS patients (25.9%) compared to non-IS patients (52.3%) developed moderate-severe ARDS, in both unadjusted (0.32; 95% CI, 0.13-0.83; p = .017) and adjusted (aOR, 0.25; 95% CI, 0.08-0.80; p = .019) analyses. After stratifying by pathologies, only IS patients with autoimmune diseases remained significant (aOR 0.25; 95% CI, 0.07-0.98; p = .046). Nonsignificant trends toward a longer time to moderate or severe ARDS, a lower need for MV/NIV, and a lower risk of death or MV/NIV were detected among IS. In our cohort of COVID-19 patients, nonsevere immunosuppression was associated with a lower risk of moderate-severe ARDS, especially among AD. This suggests a potential protective effect from a hypothesized hyper-inflammatory response.

Malignant Left Atrial Appendage Morphology: Current Classification vs H-L System.

INTRODUCTION: In previous studies the risk of stroke recurrence has been associated with the left atrial appendage (LAA) morphology (non-chicken wing (NCW)), knowing those with a greater risk as malignant LAA. Recently, a simpler morphological classification has been suggested with two categories: Low-risk (LAA-L) and High-risk (LAA-H); which could be easier to apply and may correlate better with the risk of embolic stroke. METHODS: Retrospective analysis from a registry of patients with recurrent cardioembolic strokes despite appropriate anticoagulant therapy, in which LAA morphology was studied with cardiac CT scan. LAA morphology was classified according to the four current categories and H-L morphology by the same cardiologist. Other variables associated with a high risk of stroke were also assessed, such as CHA2DS2-VASc score and left atrial (LA) size. RESULTS: Twenty-six cases were included in the analysis. We identified 22 (84.6%) chicken wing (CW), 1 (3.8%) windsock and 3 (11.5%) cactus by the current classification system, while 15 (57.7%) were classified as LAA-H and 11 (42.3%) as LAA-L by the new system. Half of the 22 cases with CW morphology were considered LAA-H, whereas all NCW were also classified as LAA-H. LA diameter and area were significantly higher in cases with LAA-H morphology (p=0.03 and 0.014), and also in those CW and LAA-H, compared to those CW with LAA-L (p=0.035). CONCLUSIONS: With this new classification system more than half of the cases of our malignant LAAs were classified as high-risk morphology. This morphology was also associated with an increased LA size.

MicroRNA-223 is a novel negative regulator of HSP90B1 in CLL.

BACKGROUND: MicroRNAs are known to inhibit gene expression by binding to the 3'UTR of the target transcript. Downregulation of miR-223 has been recently reported to have prognostic significance in CLL. However, there is no evidence of the pathogenetic mechanism of this miRNA in CLL patients. METHODS: By applying next-generation sequencing techniques we have detected a common polymorphism (rs2307842), in 24% of CLL patients, which disrupts the binding site for miR-223 in HSP90B1 3'UTR. We investigated whether miR-223 directly targets HSP90B1 through luciferase assays and ectopic expression of miR-223. Quantitative real-time polymerase chain reaction and western blot were used to determine HSP90B1 expression in CLL patients. The relationship between rs2307842 status, HSP90B1 expression and clinico-biological data were assessed. RESULTS: HSP90B1 is a direct target for miR-223 by interaction with the putative miR-223 binding site. The analysis in paired samples (CD19+ fraction cell and non-CD19+ fraction cell) showed that the presence of rs2307842 and IGHV unmutated genes determined HSP90B1 overexpression in B lymphocytes from CLL patients. These results were confirmed at the protein level by western blot. Of note, HSP90B1 overexpression was independently predictive of shorter time to the first therapy in CLL patients. By contrast, the presence of rs2307842 was not related to the outcome. CONCLUSIONS: HSP90B1 is a direct target gene of miR-223. Our results provide a plausible explanation of why CLL patients harboring miR-223 downregulation are associated with a poor outcome, pointing out HSP90B1 as a new pathogenic mechanism in CLL and a promising therapeutic target.

Genomic analysis of clonal eosinophils by CGH arrays reveals new genetic regions involved in chronic eosinophilia.

To assess the presence of genetic imbalances in patients with myeloproliferative neoplasms (MPNs), 38 patients with chronic eosinophilia were studied by array comparative genomic hybridization (aCGH): seven had chronic myelogenous leukaemia (CML), BCR-ABL1 positive, nine patients had myeloproliferative neoplasia Ph- (MPN-Ph-), three had a myeloid neoplasm associated with a PDGFRA rearrangement, and the remaining two cases were Lymphoproliferative T neoplasms associated with eosinophilia. In addition, 17 patients had a secondary eosinophilia and were used as controls. Eosinophilic enrichment was carried out in all cases. Genomic imbalances were found in 76% of all MPN patients. Losses on 20q were the most frequent genetic abnormality in MPNs (32%), affected the three types of MPN studied. This study also found losses at 11q13.3 in 26% of patients with MPN-Ph- and in 19p13.11 in two of the three patients with an MPN associated with a PDGFRA rearrangement. In addition, 29% of patients with CML had losses on 8q24. In summary, aCGH revealed clonality in eosinophils in most MPNs, suggesting that it could be a useful technique for defining clonality in these diseases. The presence of genetic losses in new regions could provide new insights into the knowledge of these MPN associated with eosinophilia.

Phosphorylation of cytosolic hPGK1 affects protein stability and ligand binding: implications for its subcellular targeting in cancer.

Human phosphoglycerate kinase 1(hPGK1) is a key glycolytic enzyme that regulates the balance between ADP and ATP concentrations inside the cell. Phosphorylation of hPGK1 at S203 and S256 has been associated with enzyme import from the cytosol to the mitochondria and the nucleus respectively. These changes in subcellular locations drive tumorigenesis and are likely associated with site-specific changes in protein stability. In this work, we investigate the effects of site-specific phosphorylation on thermal and kinetic stability and protein structural dynamics by hydrogen-deuterium exchange (HDX) and molecular dynamics (MD) simulations. We also investigate the binding of 3-phosphoglycerate and Mg-ADP using these approaches. We show that the phosphomimetic mutation S256D reduces hPGK1 kinetic stability by 50-fold, with no effect of the mutation S203D. Calorimetric studies of ligand binding show a large decrease in affinity for Mg-ADP in the S256D variant, whereas Mg-ADP binding to the WT and S203D can be accurately investigated using protein kinetic stability and binding thermodynamic models. HDX and MD simulations confirmed the destabilization caused by the mutation S256D (with some long-range effects on stability) and its reduced affinity for Mg-ADP due to the strong destabilization of its binding site (particularly in the apo-state). Our research provides evidence suggesting that modifications in protein stability could potentially enhance the translocation of hPGK1 to the nucleus in cancer. While the structural and energetic basis of its mitochondrial import remain unknown.

Effect of alemtuzumab over sNfL and sGFAP levels in multiple sclerosis.

Introduction: Alemtuzumab is a highly effective pulsed immune reconstitution therapy for multiple sclerosis (MS). Aim: To evaluate serum neurofilament light chain (sNfL) and serum glial fibrillary acidic protein (sGFAP) in patients with relapsing-remitting MS who have been treated with Alemtuzumab over the course of 2 years. Methods: This prospective study involved MS patients treated with Alemtuzumab at a referral MS center. Both sNfL and sGFAP were analyzed at baseline and then again at 6, 12, and 24 months post-treatment using the single molecule array (SiMoA) technique. We also recruited matched healthy controls (HCs) for comparison. Results: The study included 46 patients (with a median age of 34.2 [Interquartile range (IQR), 28.7-42.3] years, 27 of which were women [58%]) and 76 HCs. No differences in demographic characteristics were observed between patients and HC. The median disease duration was 6.22 (IQR, 1.56-10.13) years. The median annualized relapse rate before treatment was 2 (IQR, 1-3). At baseline, sNfL and sGFAP levels were higher in MS patients (median of 18.8 [IQR, 10.7-52.7] pg/ml and 158.9 [IQR, 126.9-255.5] pg/ml, respectively) when compared to HC (6.11 [IQR, 2.03-8.54] pg/ml and 91.0 [72.6-109] pg/ml, respectively) (p<0.001 for both comparisons). The data indicates that 80% of patients had high (≥10 pg/ml) sNfL values at baseline. We observed a significant decrease in sNfL levels at 6 (65%, p = 0.02), 12 (70.8%, p<0.001), and 24 (78.1%, p<0.001) months. sNfL reached similar levels to HC only after 24 months of Alemtuzumab treatment. During the follow-up period, no changes were identified in the sGFAP values. Conclusion: Alemtuzumab leads to the normalization of sNfL values in MS patients after 2 years of treatment, with no apparent effect on sGFAP values.

Insights into the pathogenesis of primary hyperoxaluria type I from the structural dynamics of alanine:glyoxylate aminotransferase variants.

Primary hyperoxaluria type I (PH1) is caused by deficient alanine:glyoxylate aminotransferase (AGT) activity. PH1-causing mutations in AGT lead to protein mistargeting and aggregation. Here, we use hydrogen-deuterium exchange (HDX) to characterize the wild-type (WT), the LM (a polymorphism frequent in PH1 patients) and the LM G170R (the most common mutation in PH1) variants of AGT. We provide the first experimental analysis of AGT structural dynamics, showing that stability is heterogeneous in the native state and providing a blueprint for frustrated regions with potentially functional relevance. The LM and LM G170R variants only show local destabilization. Enzymatic transamination of the pyridoxal 5-phosphate cofactor bound to AGT hardly affects stability. Our study, thus, supports that AGT misfolding is not caused by dramatic effects on structural dynamics.

Real-World Retrospective Analysis of Alemtuzumab Outcomes in Relapsing-Remitting Multiple Sclerosis: The LEMCAM Study.

BACKGROUND: Alemtuzumab is a high-efficacy treatment approved for relapsing-remitting multiple sclerosis (RRMS). Although clinical trials and observational studies are consistent in showing its efficacy and manageable safety profile, further studies under clinical practice conditions are needed to further support its clinical use. OBJECTIVE: The aim of this observational retrospective study was to evaluate the effectiveness and safety of alemtuzumab to add to the current real-world evidence on the drug. METHODS: A cohort of 115 adult patients with RRMS treated with alemtuzumab between 2014 and 2020 was retrospectively followed up in five centers in Spain. Analysis included annualized relapse rate (ARR), 6-month confirmed disability worsening (CDW), 6-month confirmed disability improvement (CDI), radiological activity, no evidence of disease activity (NEDA-3), and safety signals. Given the different follow-up periods among participants, ARR was calculated using the person-years method. CDI was defined as a ≥ 1.0-point decrease in Expanded Disability Status Scale (EDSS) score assessed in patients with a baseline EDSS score ≥ 2.0 confirmed 6 months apart. CDW was defined as a ≥ 1.0-point increase in EDSS score assessed in patients with a baseline EDSS score ≥ 1.0 (≥ 1.5 if baseline EDSS = 0), confirmed 6 months apart. RESULTS: ARR decreased from 1.9 (95% confidence interval 1.60-2.33) in the year prior to alemtuzumab initiation to 0.28 (0.17-0.37) after 1 year of treatment (87% reduction), and to 0.22 (0.13-0.35) after the second year. Over the entire follow-up period, ARR was 0.24 (0.18-0.30). At year 1, 75% of patients showed no signs of magnetic resonance imaging (MRI) activity and 70% at year 5. One percent of patients experienced 6-month CDW at year 1, 2.6% at year 2, 7.4% at year 3, and no patients over years 4 and 5. A total of 7.7% of patients achieved 6-month CDI in year 1, 3.6% in year 2, and maintained it at years 3 and 4. Most patients achieved annual NEDA-3: year 1, 72%; year 2, 79%; year 3, 80%; year 4, 89%; year 5, 75%. Infusion-related reactions were observed in 95% of patients and infections in 74%. Thyroid disorders occurred in 30% of patients, and only three patients developed immune thrombocytopenia. No cases of progressive multifocal leukoencephalopathy were reported. CONCLUSIONS: This study shows that alemtuzumab reduced the relapse rate and disability worsening in real-world clinical practice, with many patients achieving and sustaining NEDA-3 over time. The safety profile of alemtuzumab was consistent with previous findings, and no new or unexpected safety signals were observed. As this was an observational and retrospective study, the main limitation of not having all variables comprehensively available for all patients should be considered when interpreting results.

Passive assessment of tapping speed through smartphone is useful for monitoring multiple sclerosis.

INTRODUCTION: Continuously acquired smartphone keyboard interactions may be useful to monitor progression in multiple sclerosis (MS). We aimed to study the correlation between tapping speed (TS), measured as keys/s, and baseline disability scales in patients with MS. METHODS: Single-center prospective study in patients with MS. We passively assessed TS during first week, measured by an "in house" smartphone application. Reliability was assessed by intraclass correlation coefficient (ICC). Correlations between median and maximum keys/s of first week of assessment and baseline disability measures were explored. RESULTS: One-hundred three patients were included: 62.1 % women, with a median (IQR) age of 47 (40.4-54.8) years-old and an EDSS score of 3.0 (2.0-4.0). Distribution by MS subtypes was: 77.7 % relapsing-remitting MS (RRMS), 17.5 % secondary-progressive MS (SPMS) and 4.9 % primary-progressive MS (PPMS). ICC during first week was 0.714 (p < 0.00001). Both median and maximum keys/s showed a negative correlation with Expanded Disability Status Score, 9-hole peg test and timed 25-foot walk and a positive correlation with Processing Speed Test CogEval® raw and Z-score. Median and maximum keys/s were lower in patients diagnosed with SPMS than in RRMS. Both measures of tapping speed were associated with MS phenotype independently of age. CONCLUSION: TS measured through our application is reliable and correlates with baseline disability scales.

COVID-19 vaccines are not associated with axonal injury in patients with multiple sclerosis.

Objective: To evaluate the safety of COVID-19 vaccines in patients with multiple sclerosis (MS) by assessing their impact on serum neurofilament light chain (sNfL) levels as a marker of neuroaxonal damage. Methods: Single-center observational longitudinal study including patients with MS who consecutively received their initial vaccination against SARS-CoV-2 at Hospital Universitario Ramón y Cajal, following the first national immunization program in Spain. Serum samples were collected at baseline and after receiving the second dose of the vaccine. sNfL levels were quantified using the single molecule array (SIMOA) technique. Adverse events, including clinical or radiological reactivation of the disease, were recorded. Results: Fifty-two patients were included (median age, 39.7 years [range, 22.5-63.3]; 71.2% female). After SARS-CoV-2 vaccination, no increased inflammatory activity, either determined by the presence of relapses and/or new MRI lesions and/or high sNfL levels, was detected. Accordingly, there was no difference between median sNfL levels before and after vaccination (5.39 vs. 5.76 pg/ml, p=0.6). Despite this, when looking at baseline patient characteristics before vaccination, younger age associated with disease activity after vaccination (OR 0.87, 95% CI: 0.77-0.98, p=0.022). Larger studies are needed to validate these results. Conclusion: COVID-19 vaccines did not cause reactivation of disease at a clinical, radiological or molecular level, thus suggesting that they are safe in MS patients.

Structural dynamics at the active site of the cancer-associated flavoenzyme NQO1 probed by chemical modification with PMSF.

A large conformational heterogeneity of human NAD(P)H:quinone oxidoreductase 1 (NQO1), a flavoprotein associated with various human diseases, has been observed to occur in the catalytic site of the enzyme. Here, we report the X-ray structure of NQO1 with phenylmethylsulfonyl fluoride (PMSF) at 1.6 Å resolution. Activity assays confirmed that, despite being covalently bound to the Tyr128 residue at the catalytic site, PMSF did not abolish NQO1 activity. This may indicate that the PMSF molecule does not reduce the high flexibility of Tyr128, thus allowing NADH and DCPIP substrates to bind to the enzyme. Our results show that targeting Tyr128, a key residue in NQO1 function, with small covalently bound molecules could possibly not be a good drug discovery strategy to inhibit this enzyme.

Predictive models of multiple sclerosis-related cognitive performance using routine clinical practice predictors.

BACKGROUND: The application of machine learning (ML) to predict cognitive evolution is exceptionally scarce. Computer-based self-administered cognitive tests provide the opportunity to set up large longitudinal datasets to aid in developing ML prediction models of risk for Multiple Sclerosis-related cognitive decline. OBJECTIVE: to analyze to what extent clinically feasible models can be built with standard clinical practice features and subsequently used for reliable prediction of cognitive evolution. METHODS: This prospective longitudinal study includes 1184 people with MS who received a Processing Speed (PS) evaluation at 12 months of follow-up measured by the iPad®-based Processing Speed Test (PST). Six of the most potent classification models built with routine clinical practice features were trained and tested to predict the 12-month patient class label (PST worsening (PSTw) versus PST stable). A rigorous scheme of all the preprocessing steps run to obtain reliable generalization performance is detailed. RESULTS: Based on a 12-month reduction of 10% of the PST raw score, 187/1184 (15.8%) people with MS were classified as PSTw. The trees-based models (random forest and the eXtreme Gradient Boosting) achieved the best performance, with an area under the receiver operating characteristic curve (AUC) of 0.90 and 0.89, respectively. The timing of high-efficacy disease-modifying therapies (heDMTs) was identified as one of the top importance predictors in all the models evaluated. CONCLUSION: Using trees-based machine learning models to predict individual future information processing speed deterioration in MS could become a reality in clinical practice.

Counterintuitive structural and functional effects due to naturally occurring mutations targeting the active site of the disease-associated NQO1 enzyme.

Our knowledge on the genetic diversity of the human genome is exponentially growing. However, our capacity to establish genotype-phenotype correlations on a large scale requires a combination of detailed experimental and computational work. This is a remarkable task in human proteins which are typically multifunctional and structurally complex. In addition, mutations often prevent the determination of mutant high-resolution structures by X-ray crystallography. We have characterized here the effects of five mutations in the active site of the disease-associated NQO1 protein, which are found either in cancer cell lines or in massive exome sequencing analysis in human population. Using a combination of H/D exchange, rapid-flow enzyme kinetics, binding energetics and conformational stability, we show that mutations in both sets may cause counterintuitive functional effects that are explained well by their effects on local stability regarding different functional features. Importantly, mutations predicted to be highly deleterious (even those affecting the same protein residue) may cause mild to catastrophic effects on protein function. These functional effects are not well explained by current predictive bioinformatic tools and evolutionary models that account for site conservation and physicochemical changes upon mutation. Our study also reinforces the notion that naturally occurring mutations not identified as disease-associated can be highly deleterious. Our approach, combining protein biophysics and structural biology tools, is readily accessible to broadly increase our understanding of genotype-phenotype correlations and to improve predictive computational tools aimed at distinguishing disease-prone against neutral missense variants in the human genome.