Fingerstick blood assay maps real-world NAD+ disparity across gender and age

Measuring blood NAD⁺ using recombinant bioluminescent sensor protein

For achieving low-cost and rapid NAD⁺ measurement in clinical samples, we developed an NAD⁺ assay using a simple, automated reader and a fully genetically encoded bioluminescent NAD⁺ sensor protein named NS-Goji 1.3. The recombinantly expressed sensor contains an engineered NAD⁺-binding domain, a circularly permuted luciferase cpNLuc and a red fluorescent protein (RFP) mScarlet-I (Fig. 1A). NS-Goji 1.3 undergoes considerable conformational changes upon NAD⁺ binding and shortens the distance between cpNLuc and mScarlet-I, thus affecting the BRET efficiency between the pair (Fig. 1B). The highly specific NAD⁺-dependent shift of the sensor’s emission intensity at 460 nm and 580 nm is used for quantifying the NAD⁺ levels in the samples (Fig. 1C). To automatize the assay, we developed a reader that integrates a single-channel pipette robot, a photon-detector, and a reagent-containing test strip. As the E. Coli expression system makes the sensor production cost-effective and scalable, the NAD⁺ sensor is easily available for large-scale sample analysis. In parallel, an HPLC-MS quantification procedure for NAD⁺ has been developed as the reference method in which NAD⁺ was identified by ion pair of 662.0/539.7 and 80.0 m/z and quantified by the intensity of ion 662.0 m/z (Fig. 1D, 1E, and 1F).

• Download figure
• Open in new tab

Figure 1. Quantification of whole blood NAD⁺ by sensor and LC-MS.

(a) Sensing mechanism of the recombinant bioluminescent NAD⁺ sensor protein NS-Goji 1.3: The sensor changes conformation as a function of NAD⁺ concentration, leading to changes in the BRET efficiency between luciferase NLuc and RFP m-Scarlet I. (b) Sensor’s emission spectra at various NAD⁺ concentrations. (c) Ratios between sensor’s emission intensities at 580 nm and 460 nm plotted against various NAD⁺ concentrations. Values are given as mean ± SD of three independent measurements. (d) Sample preparation procedure for NAD⁺ analyzer and HPLC-MS quantification. (e) HPLC-MS retention time and ionization intensity of ions pairs for NAD⁺ identification and quantification. (f) Calibration curve for NAD⁺ measurement by HPLC-MS.

Sports and oral NMN significantly increase venous NAD⁺

Using the analyzer, we quantified the NAD⁺ levels in venous blood obtained from clinical studies to evaluate the effect of NMN supplementation on subjects between 55-70 years of age (Table S1). The double-blinded clinical test compared one placebo group and two treatment groups using different dosages of NMN (500 mg/day and 1000 mg/day) (Fig. 2A). After 30 days of daily oral administration, venous blood was collected for assessing whole-blood NAD⁺ content using both the sensor and HPLC-MS.

• Download figure
• Open in new tab

Figure 2. Measurement of venous NAD⁺ levels in clinical studies of NMN supplementation and sport.

(a) Study design for evaluating effects of oral NMN supplementation. (b) Comparison of NAD⁺ levels measured by sensor protein and HPLC-MS. (c) Bland-Altman analysis for venous NAD⁺ measured by sensor and HPLC-MS. (d) Effect of oral NMN supplementation on venous NAD⁺ level. Daily supplementation of 500 mg (n = 19) and 1000 mg NMN (n = 22) for 1 month significantly increased the venous NAD⁺ concentration compared to the placebo group (n = 25). (e) Study design for evaluating effects of oral NMN supplementation in combination with moderate level of sport. (f) Comparison of NAD⁺ levels measured by sensor protein and HPLC-MS. (g) Bland-Altman analysis for venous NAD⁺ measured by sensor and HPLC-MS. (h) Effect of sport with oral NMN supplementation (n = 20) and sport alone (n = 21) on venous NAD⁺ level. Sports significantly increased venous NAD⁺ levels in both placebo and NMN-treated groups. In (b) and (f), values are given as mean ± SD of three independent measurements. In (d) and (h), error bars represent SD of the respective group. Significance was determined using one-way ANOVA analysis for (d) and two-way ANOVA analysis for (h), * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

The measurement revealed that both dosages of NMN significantly increased the venous NAD⁺ in a dose-dependent manner. The recorded average NAD⁺ level was 23.8 ± 5.5 μM in the placebo group, 41.7 ± 13.0 μM in the 500 mg/day group and 58.8 ± 21.1 μM in the 1000 mg/day group. Notably, the variation of NAD⁺ level in the placebo group is smaller compared to that of the two treatment groups. The group treated with 1000 mg/day oral NMN featured a wide distribution of the NAD⁺ levels (SD = 21.1), indicating the possible non-responses towards the NMN administration. In addition, the comparison between NS-Goji 1.3 and HPLC-MS demonstrated a good agreement between the two methods with Pearson’s r = 0.992 (Fig. 2B, and 2C). In the subsequent clinical study, we evaluated the effect of sports on the venous NAD⁺ levels of the placebo- and NMN-treated (500 mg/day) subjects for 30 days (Fig. 2E, Table S2). NS-Goji 1.3 achieved a similar level of quantification compared to HPLC-MS for measuring venous NAD⁺ with Pearson’s r = 0.985 (Fig. 2F, and 2G). The blood NAD⁺ test showed that the moderate-intensity physical exercise significantly increased the NAD⁺ levels of both the placebo- and NMN-treated subjects compared to the control, indicating frequent exercise as a potential strategy to maintain NAD⁺ levels and to enhance the response towards NAD⁺ supplementation (Fig. 2H).

In addition to measuring NAD⁺ in blood, we explored the possibility of NAD⁺ measurement in saliva. Recent work for identifying salivary aging biomarkers revealed that salivary NAD⁺ decreases with age³¹. We hence tested if the bioluminescent sensor could quantify NAD⁺ in salivary samples. The analysis performed in parallel with HPLC-MS demonstrated that the bioluminescent sensor is capable of testing salivary NAD⁺ with Pearson’s r = 0.999 (Fig. S1). Overall, we showed that NS-Goji 1.3 sensor and the automated analyzer are viable tools for scalable and low-cost measurements of NAD⁺. The clinical study further demonstrated that regular sports enhanced the NAD⁺ in both placebo- and NMN-treated subjects.

Fingerstick blood NAD⁺ survey reveals gender-associated differences in NAD⁺ levels

For conventional NAD⁺ quantifications, venipuncture limits the frequency of NAD⁺ sampling due to invasiveness. By contrast, collecting fingerstick capillary blood by disposable lancet (Fig. 3A) would largely facilitate the dissemination of NAD⁺ measurement, which is now made possible by the bioluminescent sensor. To evaluate if fingertip blood is a representative sample source for quantifying NAD⁺, we compared the venous and fingerstick capillary blood obtained in parallel from 13 volunteers within 30 min. The NAD⁺ measured from the two sample types showed a good correlation (Pearson’s r = 0.987) and are within ± 15% error (Fig. 3B and Fig. S2). In this test, NS-Goji 1.3 measurement typically requires 5 μL of fingertip blood, while by contrast, a single venipuncture collection requires several milliliters of sample.

• Download figure
• Open in new tab

Figure 3. Fingertip blood NAD⁺ measurement for mapping NAD⁺ disparity.

(a) Procedures of taking fingerstick capillary blood for NAD⁺ measurement. (b) Comparison between NAD⁺ levels measured from fingertip and venous blood. The Black dashed line indicates y = x, and the gray dashed line indicates ± 15% deviation from y = x. (c) Study design for mapping fingertip blood NAD⁺ levels for healthy volunteers with different age and gender. (d) Female participants (n = 75) showed a significantly lower NAD⁺ level than males (n = 67) in this study. (e) 50-85 age group has significantly lower NAD⁺ levels than 18-50 age group for both male and female. The male has higher NAD⁺ level than female for 18-50 age group but not for the 50-85 age group. Significance was determined using t-test for (d) and two-way ANOVA analysis for (e), * p < 0.05, ** p < 0.01, **** p < 0.0001.

Using this fingerstick blood NAD⁺ assay, we further mapped the NAD⁺ levels of NAD⁺ supplement-naïve people across different age groups (Table S3). The cross-gender comparison indicated that females have on average a lower NAD⁺ level (27.2 μM) compared to males (32.5 μM, Fig. 3C). The survey further revealed that participants between 18-50 years of age have significantly higher NAD⁺ content than those aged between 50-85 (Fig. 3D) for both genders. Notably, the age-related NAD⁺ decrease is more pronounced for males with the average NAD⁺ level decreased from 44.2 μM to 25.9 μM (P < 0.0001). While for females, the NAD⁺ content is decreased by a lesser extent from 32.7 μM to 24.8 μM with P < 0.05. The gender-related NAD⁺ disparity is more pronounced for participants younger than 50 years old (Fig. 3D). But no significant gender-related differences were observed for participants aged over 50 (Fig. 3D). Interestingly, a few participants have particularly high NAD⁺ levels for the respective age group, indicating the existence of outliers whose NAD⁺ metabolism is markedly different. Further studies to identify the genetic and/or acquired factors for this high NAD⁺ level should provide more insight for developing new NAD⁺ modulating strategies.

Measuring real-world NAD⁺ dynamics using minimal capillary blood samples

NAD⁺ metabolism plays a central role in Circadian rhythm. However, the hypothetical human NAD⁺ differences between day and night have not been recorded due to the difficulties of collecting venous blood after midnight. As fingertip blood assay is easily applicable at the bedside, we collected and measured the capillary blood NAD⁺ at 4 am and 10 am from n = 7 subjects (Fig. 4A). The measurement showed that the capillary NAD⁺ levels did not change significantly between 4 am at 10 am (P = 0.0735, Fig. 4B). To evaluate the potential fluctuation in NAD⁺ during the day, we further accessed the capillary NAD⁺ levels after oral administration of 300 mg of NMN during the day starting from 10 am. The NAD⁺ level significantly increased 60 min after the NMN administration and returned quickly to the basal level at 120 min (Fig. 4C). Apart from the short NMN-induced perturbation, no significant changes were observed during the day, demonstrating the fast metabolism of NMN and the strong homeostasis of the NAD⁺ level. To further map the NAD⁺ fluctuations in the real world over long terms, we measured the NAD⁺ levels of n = 6 subjects twice a week over 100 days and recorded major changes in dietary, sportive, and sleeping habits (Fig. 4D). The result showed that NAD⁺ levels were mostly stable during the 100-day monitoring (Fig. 4E&F), except for one participant who spontaneously started NMN administration which induced a notable increase in the capillary NAD⁺ level (from 32.9 to 47.3 with p < 0.0001). One subject exercised regularly (11 out of 25 days on which NAD⁺ was measured), and the recorded NAD⁺ levels on days with sport were significantly higher than days without. However, the sport-induced NAD⁺ increase was not observed for the subject with a low sport frequency (3 out of 24 days of NAD⁺ measurement). Other recorded events such as sleep deprivation, menstruation, oral administration of phosphatidylcholine (PPC) and silibinin did not significantly affect the capillary NAD⁺ level (Fig. 4G).

• Download figure
• Open in new tab

Figure 4. Monitoring real-world NAD⁺ dynamics using capillary blood.

(a) Study design for comparing capillary NAD⁺ levels at 4 and 10 a.m. (b) No significant difference in NAD⁺ level was observed for capillary blood samples obtained at 4 and 10 a.m. (n = 7, p = 0.0735), though the average NAD⁺ level increased from 36.00 μM to 38.69 μM. (c) Oral administration of 300 mg NMN induced short-term capillary NAD⁺ spike with significant increases recorded at 60 min (n = 5, p = 0.0034). (d) Scheme for the long-term monitoring of NAD⁺ levels. Events including sport, sleep shortage, medication, menstruation, and dietary supplementation were recorded through questionnaire. (e) NAD⁺ dynamics recorded for n = 6 participants over three months. F and M indicates female and male subjects respectively. (f) Distribution of fingerstick NAD⁺ levels during the long-term monitoring. (g) Effects of events on NAD⁺ levels for each subject. NMN and sport significantly increased NAD⁺ levels for sub 3 and 6 respectively. Only days with single recorded event were considered, while days with more than one recorded event were excluded from the analysis. Significance was determined using t-test for (c) and sub 4 & 6 in (g). Two-way ANOVA analysis was used for significance determination for sub 1, 2, 3 and 5 in (g), * p < 0.05, **** p < 0.0001.

Taking together, the capillary blood NAD⁺ measurement revealed significant NAD⁺ disparity across gender and age. The sensor also facilitated the real-world observation of NAD⁺ modulation by dietary supplements and sports. We further demonstrated that individual NAD⁺ levels are relatively stable over a long period of time and continuous oral NMN administration and frequent sport can boost NAD⁺ levels. This easy-to-implement NAD⁺ test for blood and saliva provided a useful tool for monitoring the effect of various NAD⁺ modulating strategies at the point-of-care.

Clinical subjects and samples

Study protocol for evaluating the effect of oral NMN supplementation and sports on human subjects have been reviewed and approved by the ethics committee of Guangzhou Sport University (ethics approval number: 2020 LCLL-003) in accordance with the Declaration of Helsinki. First, the study evaluates how different doses of NMN affect the blood NAD⁺ level. Then, the study compares the effect of sports with and without NMN supplementation on the blood NAD⁺. The study recruited participants between 55-70 years of age. The exclusion criteria are 1) having hepatic and/or renal dysfunction, infectious diseases, cardiac dysfunction, and/or spinal dysfunction, 2) having chest pain or shortness of breath during sport, 3) having more than 3% changes of body weight within 2 months before the study, 4) having previously taken nicotinic acid, nicotinamide mononucleotide, nicotinamide, or other vitamins B3-related dietary supplements, 5) taking coffee or caffeine-containing drinks, 6) during pregnancy or lactation, 7) being mentally unfit to participate in the study. For the sport intervention, subjects participated in centralized aerobic sport for 40-60 min, three times per week. For the first two weeks, the sport intensity was maintained between 40% to 59% of the heart rate reserve (HRR), then sport intensity was increased to 60% to 85% of the HRR. In addition, subjects participated in resistance training 3 times per week. For the dietary supplementation, subjects were grouped and tested in a double-blinded, placebo-controlled manner with three groups taking the placebo, 500 mg/day NMN, and 1000 mg/day NMN.

The study protocol for evaluating fingertip blood NAD⁺ levels of human subjects have been reviewed and approved by the Institutional Review Board of Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences (ethics approval number: SIAT-IRB-210915-H0575) in accordance with the Declaration of Helsinki. Fingerstick sampling was performed to collect 5 μL of capillary blood using disposable fingerstick lancets (STERiLANCE Medical Inc.) for NAD⁺ measurement. The blood sampling time, administration of medication and/or dietary supplements, physical exercise, sleep quality and menstruation have been recorded for data analysis with written consent of participants. The exclusion criteria are 1) being younger than 18 years of age, and 2) having hemophobia or other symptoms hindering the fingerstick blood sampling.

Sensor preparation

The gene of the sensor NS-Goji 1.3 was constructed in a pET15b (+) vector for recombinant expression in E. coli (DE3). The E. coli (DE3) was cultured 1 L LB medium containing 50 μg/mL ampicillin at 37 °C with a shaking rate of 220 rpm to reach an optical density of 0.6-0.8 at 600 nm. Then 0.5 mM isopropyl β-D-thiogalactopyranoside (IPTG) was added into the medium to induce sensor expression at 16 °C for 16 hours. The E. coli (DE3) was collected by centrifugation at 8000 rpm for 15 min and the resulting pellet was suspended in 25 mM Tris-HCl buffer containing 500 mM NaCl and 20 mM imidazole at pH 8.0 for lysis using high-pressure homogenizer at 4 °C. The lysate was centrifuged at 13,000 g for 30 min at 4 °C and the resulting supernatant was filtrated by 0.22 μm syringe filters for purification by Ni-NTA (Smart Lifesciences Lnc., China) and Strep-Tactin (Smart Lifesciences Lnc., China) columns. The purified protein was desalted and exchanged into 50 mM HEPES buffer containing 50 mM NaCl (pH 7.2) with an Amicon Ultra-15 centrifugal filter (10 kDa MWCO, Merk Millipore Inc.). Protein concentration was measured by Bradford assay. The obtained protein solution was added into an equal volume of glycerol for long-term storage at −80 °C.

Blood sample measurement by NAD⁺ sensor

Phlebotomy was performed to collect venous blood samples in plastic blood collection tubes with EDTA as anticoagulant. Fingerstick sampling was performed to collect capillary blood samples using disposable fingerstick lancets (STERiLANCE Medical Inc.). The blood samples were processed by the analyzer’s automated sample handling program in which 5 μL of sample were lysed by mixing thoroughly with 4 volumes of 0.5 N perchloric acid. The lysate was then diluted by 10 folds in a buffer containing 500 mM HEPES, 500 mM NaCl at pH 7.2. 10 μL of the resulting solution was then added into 90 μL of measurement buffer containing 0.5 nM sensor, 100-fold diluted furimazine, 50 mM HEPES and 50 mM NaCl at pH 7.2, which was previously lyophilized and automatically reconstituted in the analyzer’s mixing well by the liquid handler. The sensor’s bioluminescent signal was measured using the integrated photon-detector (Hamamatsu Photonics) of the analyzer with NLuc emission measured at 460 nm (bandwidth 30 nm) and RFP emission was measured at 580 nm (bandwidth 20 nm). The sensors’ emission ratio (R) between 580 nm and 460 nm was used to calculate the NAD⁺ concentrations according to the titration curve.

To titrate the senor, standard aqueous NAD⁺ solutions with known concentrations were prepared and measured by the above-mentioned method. The sensors’ emission ratio (R) was plotted against the NAD⁺ concentrations and the resulting curve was fitted to the Hill-Langmuir equation to obtain the sensor’s maximum ratio (Rmax), minimum ratio (Rmin), c50 and Hill coefficient (h) NAD⁺ concentrations in unknown samples were calculated with the rearranged the equation:

Blood sample measurement by HPLC-MS

5 μL of whole blood was taken from blood collection tubes and added into 20 μL of 0.5 N perchloric acid with thorough mixing. Cell debris and aggregates were separated by centrifugation at 12000 g for 5 min at 4 °C. The supernatant was transferred into sample tubes and stored at −80 °C until analysis. Upon analysis, the supernatant was diluted by 10 folds in aqueous solution with 50% methanol and 50% water and then filtered with a PTFE syringe filter (0.45 μm * 13 mm) before injection into LC-MS.

Agilent Infinity Lab LC/MSD iQ G6160A was used to perform the LC-MS quantification of the blood samples. Sciex’s MultiQuant v.3.0.2 software was used for instrument control, data acquisition and data processing. 10 μL of each prepared sample were injected into LC and separated by a ZORBAX SB-Aq Column (Agilent, 4.6 × 100 mm, 3.5 μm). The autosampler temperature was set to 4 °C and the column temperature was set to 20 °C. The following elution program was used: A = acetonitrile, B = 0.1% formic acid-methyl alcohol, 0-0.3 min, 80% A; 0.3-1.5 min, 80-20% A; 1.5-2.5 min, 20% A; 2.5-2.6 min, 20-80% A; 2.6-3.0 min, 80% A, with total run time of 3.0 min and a flow rate of 1.00 mL/min. For MS quantification, MS was equipped with electrospray ionization (ESI) ion source and operates in negative ion mode. Standard solutions with known concentrations of NAD⁺ (0.05, 0.1, 0.2, 0.5, 1.0, 2.0 μg/mL) have been prepared and automatically injected with single injections for LC-MS analysis to optimize the quantification conditions. Precursor ions with the highest response were used as quantitative ion pairs and daughter ions with the higher abundance and less interference were used as qualitative pairs. The data analysis was performed with Sciex’s MultiQuant v.3.0.2 software. This quantification method showed a limit of detection (Signal to noise ratio ≥ 3) of 0.0200 μg/mL and a limit of quantification (Signal to noise ratio ≥ 10) of 0.0500 μg/mL.

Determination of recovery rate and sample stability

100 μL of solution was taken from lysed blood samples with known NAD⁺ concentrations. The samples were then added with 100, 400 and 700 μg/L standard NAD⁺. The samples were further diluted in aqueous solution with 50% methanol and 50% water for six parallel LC-MS measurements.

Blood samples lysed with 0.5 N perchloric acid were stored at room temperature and at 4 °C for 0 h, 1 h, 2 h, 4 h, 6 h, 12 h, 24 h, and 48h before LC-MS measurement using the above-mentioned method.

Salivary sample measurement

Salivary samples were collected in 10 mL conical tubes in the morning before breakfast after at least 8 hours of fasting, which were then added with various concentrations of NAD⁺ solutions to prepare spiked salivary samples. The resulting samples were loaded into the NAD⁺ analyzer and were automatically added with 4 volumes of 0.5 N HClO₄. The analyzer then dilutes the resulting mixture by 10 folds in a buffer containing 500 mM HEPES, 500 mM NaCl at pH 7.2, and transfers 10 μL of the resulting solution into 90 μL of measurement buffer containing 0.5 nM sensor, 100-fold diluted furimazine, 50 mM HEPES, 50 mM NaCl at pH 7.2 for bioluminescent measurement. The sensor’s bioluminescent signal was measured using the analyzer’s integrated photon-detector (Hamamatsu Photonics) with NLuc emission measured at 460 nm (bandwidth 30 nm) and RFP emission was measured at 580 nm (bandwidth 20 nm). The sensors’ emission ratio (R) between 580 nm and 460 nm was used to calculate the NAD⁺ concentrations according to the titration curve performed in parallel. HPLC-MS measurement of the supernatant was performed in parallel according to the above-mentioned method as reference.

Statistics

All values were expressed as mean ± standard deviation. Statistics were performed using Prism 8 (Graphpad). All quantification tests were independently performed for at least 3 times.

Correspondence and requests for materials should be addressed to L.C., M.H. and Q.Y..