Stability of aqueous solutions of ascorbate for basic research and for intravenous administration

Abstract

Ascorbate (vitamin C) can rapidly oxidize in many near-neutral pH, aqueous solutions. We report on the stability of ascorbate solutions prepared for infusion into patients using standard pharmacy protocols, for example, 75 g of ascorbate/L in water for infusion. The concentration of ascorbate was monitored for changes over time using direct UV–Vis spectroscopy. The pH of the solution was about 5.7 with no significant change over 24 h. There was only an approximate loss of 1% per day over the first 3 days of storage. This information allows decisions on how far ahead of need such preparations can be made. We also provide laboratory approaches to minimize or control the rate of oxidation of ascorbate solutions for use in chemical and biochemical studies as well as preclinical animal studies. The goal is to have the amount of ascorbate intended to be used in experiments be the actual amount available.

1. Introduction

Since its identification by Albert Szent-Györgyi nearly 100 years ago, vitamin C, i.e. ascorbic acid or ascorbate, has been known as an easily oxidizable reducing agent [1,2]. Its rate of oxidation is accelerated by metal ions such as copper and iron, as well as base (OH⁻), i.e. high pH [1,3,4]. Thus, to study its chemistry and to preserve it in solution, redox active metals ions must be carefully controlled, and a somewhat acidic environment is preferred [3,5].

An overview of the acid-base and redox species of ascorbic acid is presented in Fig. 1 (Modified from [6].). pK_a values shown for ascorbic acid and the ascorbate monoanion are from [7–13]. The pK_a’s of ascorbic acid indicate that at pH 7.4 in typical buffer solutions, 0.04% will be present as the diacid, 99.94% will be present as the monoanion and only about 0.02% will exist as the dianion (Asc²⁻). In this work we will typically refer to vitamin C as ascorbate (AscH⁻), because the monoanion is the dominant form in vivo and is responsible for the biochemistry that makes it a vitamin.

Fig. 1.

The various forms of ascorbate. The ascorbate monoanion (AscH⁻) is the dominant form at physiological pH. The pK_a1 of the diacid is around 4.2 at 25 °C as the ionic strength approaches 0; but at typical ionic strengths of laboratory buffers and blood plasma (≈150 – 200 mM), pK_a1 will be about 4.0, and even a bit lower at 37 °C. pK_a2 at 25 °C is around 11.6. When the ionic strength is in the range of 150 – 200 mM, the effective pK_a2 at 37 ^°C is probably around 11.2 Note that pK_a =−0.45 for the ascorbyl radical (charge neutral); thus, it is a strong acid and will not exist in any biological setting; only the ascorbate radical (Asc^•−) will exist at near-neutral pH. The species shown in blue are those that principally contribute to the basic biochemistry and biology of ascorbate. AscH₂ is the diacid, i.e. ascorbic acid; AscH⁻, the ascorbate monoanion; Asc^•−, the ascorbate free radical; DHA, the two-electron oxidation product, dehydroascorbic acid. Although the structure labeled as pseudo-DHA is typically presented as the two-electron oxidation product, it is not an important species. See Note on nomenclature and abbreviations, Section 2.7. Modified from [6].

There are two possible forms of the one-electron oxidation product of ascorbate, the ascorbyl radical (AscH^•) and ascorbate free radical (Asc^•−). The pK_a of AscH^• has been determined to be −0.45 [14], i.e. it is a strong acid. Thus, AscH^• has no role in biology; Asc^•− is the only form that will exist at biological pH values.

The biochemically active, two-electron oxidation product of AscH⁻ is dehydroascorbic acid (DHA) [15]. DHA has two fused 5-membered rings containing a fragile lactone, i.e. carboxylic acid ester Fig. 2. If DHA is not reduced back to AscH⁻, it will undergo hydrolysis resulting in ring-opening, Fig. 3 [16]. Once the ring opens, its activity as vitamin C and as a reducing agent is irreversibly lost. The hydrolysis is pH-dependent, being facilitated by OH⁻ as well as H⁺ [17].

Fig. 2.

The two-electron oxidation shown with the best accepted structures of the species.

Fig. 3.

Oxidation-reduction of ascorbate with hydration of dehydroascorbic acid (DHA) and subsequent hydrolysis. Ascorbate readily undergoes sequential one-electron oxidations to form the ascorbate free radical and then dehydroascorbic acid. Upon hydrolysis, the rings of DHA open and ascorbate is essentially lost. DHA can be reduced by glutathione in conjunction with glutaredoxin (Grx), as well as other enzymatic systems, thereby recycling ascorbate. However, if not reduced, it can further breakdown to give a variety of products. Shown here are examples. Note that protons, water, and CO₂ are not shown. Adapted from [6,28]. See Note on nomenclature and abbreviations, Section 2.7.

The primary degradation product in the circulation is 2,3-diketogulonate, Fig. 3 [18]. 2,3-Diketogulonate then can breakdown to a whole host of products with varied chemistry that can complicate basic studies of ascorbate and may lead to adverse health effects [16,19,20].

The structure labeled as pseudo-DHA in Fig. 1 may be fleeting as it would readily hydrate to form DHA; however, it may not exist at all; the analysis of Njus et al. indicates that it is energetically nearly impossible to achieve [21]. The structure labeled pseudo-DHA in Fig. 1 is widely presented in the literature as the two-electron oxidation product of ascorbic acid/ascorbate. This is probably the result of some unfamiliarity with this detail of its chemistry and under appreciation (over simplification) of the chemical intermediates involved in the oxidation of ascorbate. This incorrect chemistry has been propagated in the literature and could hinder progress [15]. We encourage the presentation of the stoichiometry and structures presented in Figs. 1, 2, and 3.

Exceptionally strong evidence has been presented that it is the fused ring, i.e. two 5-membered rings, that is actually active in vivo, and not pseudo-DHA [22]. As might be expected, the ring structure has many similarities to glucose and thus would be amenable to uptake by membrane glucose transporters, Fig. 2.

Here we present our laboratory and clinical experience in handling aqueous solutions of ascorbic acid/ascorbate that minimizes unwanted loss before use in the laboratory or in the clinic.

2. Materials and methods

2.1. Materials for laboratory use

Currently we use L(+) Ascorbic Acid Fine Crystals, USP grade, MW = 176.13 g mol⁻¹, CAS 50–81–7 from Macron Fine Chemicals, Avantor Performance Materials, Inc., Center Valley, PA, USA, or EMD Chemicals, Darmstadt, Germany.

However, other suppliers provide excellent ascorbic acid as the diacid, as white fine crystals. We have found that the ascorbic acid from some suppliers to be unsatisfactory for our goals; caution is advised. We do not use sodium ascorbate as we have found that most commercial preparations have considerable levels of impurities as judged by the off-white color of the solid and yellow color of freshly prepared aqueous solutions. These impurities are probably the downstream oxidation products of ascorbate due to its oxidation in the preparation of the sodium ascorbate.

2.2. Materials for clinical use, i.e. IV infusion into human subjects

To examine the stability of ascorbate solutions prepared for IV (intravenous) delivery, three vials of ASCOR^® Ascorbic Acid 25,000 mg/50 mL for Injection (Lot #: 19K0151) from McGuff Pharmaceuticals, Inc (Santa Ana, CA) were introduced into a PINNACLE^® Single Chamber Mixing Container, EVA Formulation, 1000 mL (B. Braun product # 2112348 [23]). Some of the water was removed so the bag contained 850 mL of Sterile Water for Injection USP (Baxter Healthcare, 3000 mL Lot # Y321158). Upon introduction of the 150 mL of ASCOR^®, the bag then contained 1000 mL of ascorbate for IV delivery, exactly parallel to what is prepared for patients (temperature of the solution was about 5 °C). The expected concentration of ascorbate in the IV bag was about 426 mM. This protocol is identical to that for preparation for IV infusion for our clinical trials examining the potential of P-AscH⁻ as an adjuvant in cancer therapy.

The package insert for ASCOR^® states: “Each mL of ASCOR contains 500 mg of ascorbic acid (equivalent to 562.5 mg of sodium ascorbate which amounts to 65 mg sodium/mL of ASCOR), 0.25 mg of edetate disodium, and water for injection. Sodium hydroxide and sodium bicarbonate are added for pH adjustment (pH range 5.6 to 6.6). It contains no bacteriostatic or antimicrobial agent.” [See ASCOR^® description at https://www.mcguffpharmaceuticals.com/ascor, as accessed on 2023.06.28.]

The final product was placed in a brown bag that protects from light and stored in a refrigerator (between 2 °C and 8 °C). The infusion bag contained about 100 mL of headspace air. Using a 2 mL syringe, samples were taken periodically for analysis of ascorbate, which would be present principally as the monoanion.

Samples from the IV bag (20 μL) were diluted into 380 μL of a 2.9 mM phosphate-buffered saline solution, pH 6.5 that had been treated and stored with Chelex^® beads to remove adventitious metals [3]. The diluted and mixed samples were then analyzed immediately by UV–Vis spectrophotometry using ≈ 1 μL samples. For the direct determination of ascorbate, absorbance measurements at 265 nm (ε₂₆₅ = 14,500 M⁻¹ cm⁻¹ [3]) were accomplished using both P-330 and C40 Implen Nano-photometers with their respective 250x dilution lids (i.e., path length = 4.00 × 10⁻³ cm = 40 μm) [24]. With dilution of the sample, as above, and a pathlength of 4.00 × 10⁻³ cm the expected absorbance would be about 1. A full spectrum was always collected to ensure there were no quality control issues. If a UV–Vis spectrometer setup is used having a different pathlength, the samples will need to be diluted appropriately.

Plasma samples were from subjects in an ongoing clinical trial (https://clinicaltrials.gov/ct2/show/NCT02905578); samples were collected and handled as described in detail in [25].

2.3. UV–Vis absorbance measurements

Ascorbic acid and the ascorbate monoanion have distinctly different ultraviolet (UV) spectra. The diprotic acid, AscH₂, has a maximum molar absorptivity at 243 nm, ε₂₄₃ = 9650 M⁻¹ cm⁻¹ [26]. The monoanion has an absorption maximum at 265 nm. Many values for its molar absorptivity have been reported, with values for ε₂₆₅ ranging from 7500 to 20,400 M⁻¹ cm⁻¹ [3,12,26]. A careful study found ε₂₆₅ = 14,560 ± 450 M⁻¹ cm⁻¹ and ε₂₅₁ = 8250 ± 150 M⁻¹ cm⁻¹ at the isosbestic point (250.7 nm) for the diacid and the monoanion [26]. When unwanted oxidation is avoided, ε₂₆₅ = 14,500 M⁻¹ cm⁻¹ is the maximum UV absorbance of AscH⁻in our hands [3]. However, the effective molar extinction coefficient of AscH⁻in human blood plasma appears to be ≈10% smaller, ε₂₆₅ = 13,000 M⁻¹ cm⁻¹ [24]. Because the ascorbate in the present study is in simple aqueous solutions, the molar extinction coefficient of 14,500 M⁻¹ cm⁻¹ was used to determine concentrations.

2.4. Buffer capacity

The buffer capacity of the ascorbate solution (75 g L⁻¹) was determined by sequential addition of 10 μL of 1.00 M HCl or 1.00 M NaOH (10 × 10⁻⁶ moles) into 8.00 mL (0.008 L) of the solution and monitoring the change in pH. This translates to 1.25 mmol L⁻¹. The buffer capacity (β) was determined using:

$$\beta =\left(n\times 1.25{\text{mmol}\text{L}}^{-1}\right)/\Delta \text{pH}(\mathit{\text{units}}\mathit{\text{are}}\mathit{\text{mM}}/1\text{pH}$$

where n = number of additions of acid or base and ΔpH is the change in pH with the addition(s). The change in volume of the solution upon sequential additions of acid or base was taken into account.

2.5. Making stock solutions of ascorbate

When handling ascorbate solutions, pH and adventitious catalytic metals are prime considerations for avoiding unplanned oxidation. Ascorbic acid is a diacid with pK_a’s of around 4.0 and 11.0 in typical buffer solutions [7,8,10–13,27,28]. Thus, at near-neutral pH the dominant form is the ascorbate monoanion, Fig. 1. Ascorbate is easily oxidized; as the pH of an aqueous solution is increased the rate of oxidation increases due to the increased concentration of the ascorbate dianion [3,10]. However, some catalytic metal salts, such as those of copper and iron that have permissive thermodynamics and coordination environments, can efficiently catalyze the oxidation of the ascorbate monoanion [1,3,5,29,30]. Even small amounts of adventitious catalytic metals present in typical buffers will lead to rates of oxidation far, far greater than the direct reaction of the ascorbate dianion with dioxygen [3].

When making stock solutions of ascorbic acid/ascorbate, we make them very concentrated, e.g. 1.0 M or 0.10 M in clean glassware using high purity water. See Note on volumetric ware, Section 2.8. For in vivo injection the stock solution is made to have a target concentration of 1.0 M. We use white, crystalline ascorbic acid, i.e. the di-acid, to make our stock solutions. We find that commercial preparations of sodium ascorbate have considerable levels of impurities as judged by the off-white color of the solid and yellow color of fresh aqueous solutions. Our absorbance measurements of carefully prepared aqueous solutions of sodium ascorbate verify that oxidation has occurred in the commercial stock. Although sodium ascorbate has its place in the laboratory, in our experience, don’t waste time and money on sodium ascorbate when rigorous and reproducible studies are planned centered on the ascorbate monoanion.

To determine the concentration of our ascorbate stock, we dilute appropriately into buffer; pH 6 – 7.5, best is pH 6 – 7. Greatly preferred is buffer that has been “de-metalled” using chelating beads employing the batch method [3]. See Note on removing adventitious catalytic metals from buffer solutions, Section 2.9. We use the absorbance at 265 nm, ε₂₆₅ = 14, 500 M⁻¹ cm⁻¹, to verify and adjust concentrations [3].

At these high concentrations, only a very small fraction of the ascorbate will oxidize [5]. The concentration of dissolved oxygen in air-saturated aqueous solution is about 250 μM at room temperature [31–33]; thus, oxygen is the limiting reagent. That is, as the ascorbate oxidizes all the dissolved oxygen in the solution will be consumed, assuming a closed vessel with no headspace air. See Note on headspace air, Section 2.10. Once the dissolved O₂ is depleted, no more oxidation of ascorbate will occur. Initially, in an air-saturated room temperature aqueous solution of 1.0 M ascorbate, the concentration of dissolved oxygen is about 0.00025 M [31]. If electrons from ascorbate fully reduce O₂ to water (2AscH⁻+ O₂ + 4H⁺ → 2H₂O + 2DHA), then 0.0005 M of ascorbate will be lost to this oxidation, or (0.0005 M/1. 0 M) x 100% = 0.05%, which is insignificant in most applications. Thus, the goals in the preparation and handling of ascorbate stock solutions are to:

1. minimize oxygen;

2. minimize catalytic metals; and

3. make a somewhat acidic stock, with the actual pH of the stock solution dependent on needs.

2.6. Detailed protocol for making stock solutions of ascorbate

1. For injection, animal studies. For in vivo injection the stock solution is made to have a target concentration of 1.00 M. We use white, crystalline ascorbic acid, i.e. the di-acid, to make our stock solutions. The stock is prepared by directly putting the gravimetric appropriate amount of solid ascorbic acid into a conditioned flask then 18 mΩ water that has been N₂- or Ar-gassed to remove most of the ambient oxygen is added to 85% of the theoretical final volume while slowly mixing with a stir bar. After it is fully dissolved, we adjust the pH with 10 mM NaOH to near pH 7.0: caution when approaching pH 6.5 or higher as the pH changes rapidly with small volumes of NaOH and can easily overshoot pH 7. Additional water and NaOH are added to achieve the final volume and pH 6.95–7.0. The concentration of the stock solution is verified spectrophotometrically and noted. This high concentration stock solution is actually self-sterilizing.

   Typically, we prepare 400 mL of 1.00 M stock pH ≈ 6.95 – 7.05 and aliquot into glass screw top tubes with as small of a volume of headspace as possible. The stock is then stored at 5 °C (refrigerator) away from light. When properly aliquoted and stored without disturbance or opening the stock is stable for several months. When prepared appropriately the resultant stock solution is clear; however, if yellowing is detected during storage the stock is discarded and a new one prepared.

2. For chemical/biochemical studies. Ascorbate stock prepared for injection in animal studies is often used for most in vivo and in vitro studies, but this really depends on the type of experiment. However, there are occasions in chemical/biochemical studies that do NOT require a pH adjusted ascorbic acid stock. In these cases a gravimetrically prepared stock solution can be made in water without adjusting the pH. The exact concentration of the stock can be confirmed with UV/Vis absorbance using ε₂₆₅ = 14,500 M⁻¹ cm⁻¹ at 265 nm after diluting appropriately into a pH ≈6.5 de-metaled buffer [3]. This stock, when sealed, and refrigerated at ≈ 5 °C away from light and is quite stable; however, we replace it if hints of yellow appear. Note, this stock solution of ascorbic acid in DI water is NOT pH adjusted. The lower pH increases its stability; see Note on DHA and stability, Section 2.11.

3. For our typical studies, immediately before use, the stock is diluted appropriately into deionized-water (or distilled water) to yield a new stock solution that is 10x, 20x, 25x or possibly 50x of the final concentration we want in our experiments. Depending on which stock solution, this approach may produce a somewhat acidic stock solution of ascorbate that is relatively stable for some time. When introduced into the experiment, the buffers present in the solutions will typically overwhelm the non-buffered ascorbic acid stock being added. There is usually little if any change in pH. This should be verified, of course.

4. The concentrated ascorbic acid stock solution keeps for some days or weeks, depending on the glassware and availability of oxygen. If there is any hint of yellow, a new stock is made using the same flask. However, any diluted stocks that have been made in DI water are replaced a few hours after making them. Compared to the overall experiment, the cost in time and resources is minimal to renew these diluted stock solutions. (For example, if we made a diluted solution of ascorbate in water in the morning, we would probably quickly make a new diluted solution for use in the afternoon.)

5. Scale up or down as appropriate for your needs.

6. For cell culture experiments we use stock solutions produced from either approach, but the stock solutions of ascorbate for injection are preferred by most. We find that moles of ascorbate per cell is the best, most reproducible, and most informative approach to specify dose of ascorbate in cell culture work [34,35]. It is far superior to nominal concentration, for example mM. In publications, we typically provide both forms of dose, i.e. mol cell⁻¹ and molar (μM, mM etc.). Most experimenters in our greater group have adopted the mol cell⁻¹ approach in their cell culture work because of its many advantages, and especially for Rigor and Reproducibility.

7. Exposure of cells to P-AscH⁻. When cells are at the desired growth state, they are then rinsed with new medium; this medium should not contain pyruvate, as it reacts with H₂O₂ introducing a complexity that is nearly impossible to resolve (See discussion of this issue in the Supplemental Information of reference [35].). Cells are treated with ascorbate (the best metric to specify dose is as pmol cell⁻¹ [35]) from a stock solution of 1.00 mol L⁻¹ ascorbate (pH 7.0), which was made under N₂ or Ar and stored in a volumetric flask with a tight-fitting stopper at 4 °C. The concentration of ascorbate in the stock solution will have been verified using its UV/Vis absorbance ε₂₆₅ = 14, 500 M⁻¹ cm⁻¹ at 265 nm after diluting appropriately into a pH ≈ 6.5 buffer [3]. See Section 2.3 above. This solution can usually be kept for several weeks without significant loss of ascorbate due to the lack of oxygen. Cells are treated with ascorbate for 1 h at 37 °C [36,37]. Then the ascorbate-containing medium is removed and cells are ready for the next phase of experimentation or assessment. For example, for a clonogenic cell survival assay, cells are then rinsed, trypsinized, counted, diluted, and plated. (See Note on exposure of cells in culture to ascorbate, Section 2.12.)

2.7. Note on nomenclature and abbreviations

We now use the abbreviations shown in Fig. 1 and Table 1 for the various chemical species the arise from vitamin C. Our use of abbreviations has evolved in an effort to better show some aspect of the chemical species as well as the chemical name. Because ascorbic acid is present as the monoanion at near-neutral pH, it is the key species in the biochemistry and subsequent biology of vitamin C. We refer to it as ascorbate, not ascorbic acid, especially when we want to emphasize the chemistry/biochemistry. For pharmacological ascorbate, i.e. the use of ascorbate as a drug, we use P-AscH to emphasize we are focused on its action as a pro-drug for the production of H₂O₂ [38]; however, its function as a vitamin is not overlooked.

Table 1.

Abbreviations used for various species arising in ascorbate chemistry.

Abbreviation	Common Name	Comment
AscH2	Ascorbic acid	the diacid
AscH−	Ascorbate monoanion	The dominant form of vitamin C in biology
Asc2−	Ascorbate dianion	A very minor species
AscH•	Ascorbyl radical (Semidehydroascorbic acid)a	Because this radical has a neutral charge the “yl” suffix is appropriate
Asc•−	Ascorbate radical (Semidehydroascorbate)a	With a negative charge the “ate” suffix is appropriate.
P-AscH−	Pharmacological ascorbate	For use when ascorbate is being used as a drug, e.g. a pro-drug [38].
Pseudo-DHA	Pseudo-dehydroascorbic acid; see [11].	Albeit the chemical structure seen in Fig. 1 is widely used in the literature, it is very minor or even non-existent form of the 2-electron oxidation product of ascorbate [15,21].
DHA	Dehydroascorbic acid	The compound with two fused 5-membered rings is the major form of the 2-electron oxidation product of ascorbate/ascorbic acid.

^aDehydro means the loss of hydrogen atoms (electrons), typically two hydrogen atoms (or an even number) are lost. Thus, “semidehydro-“ is the loss of one hydrogen atom.

These abbreviations work well as they indicate aspects of the chemistry, are unique, and are not gaudy when used in text of written work or on slides.

2.8. Note on volumetric ware

In the lab we are mindful of the volumetric ware we use because of the potential for contamination of solutions they contain, especially redox active metals, or with plasticware, the possible presence of antioxidants that can migrate into solutions [39,40]. Be aware that Pyrex glass may have a smidgen of iron in its formulation, either purposefully or adventitiously. It is our observation that this iron, probably on the surface, can catalyze the oxidation of ascorbate. Thus, before initial use of an item of glassware, we soak it in dilute acid for some time to remove some of this iron. We actually purchase a new piece of glassware, treat it for some time with dilute acid before first use. Then, we use this item of glassware for the same stock solution year after year. Some glassware in our laboratory have been in use for specific stock solutions for over two decades and are now well-conditioned. They perform superbly. Avoid use of ground glass to store solutions and be especially careful to avoid having ascorbate solutions come in contact with ground glass [41]. The large surface area may result in exposure to a great deal of iron as it is a trap for metals.

Some plasticware, especially polypropylene, can contain hindered amine antioxidants. These and their downstream oxidation products, i.e. nitroxides, will migrate into solutions providing hidden antioxidants that can change the results of free radical experiments [39,40]. Thus, in general we avoid putting solutions into polypropylene plasticware, such as syringes, and centrifuge tubes. However, rinsing syringes or tubes with water several times can reduce the levels of these compounds when intended for short-term use, i.e. non-storage.

2.9. Note on removing adventitious catalytic metals from buffer solutions

To remove adventitious catalytic metals from typical laboratory buffers, such as near-neutral phosphate buffers, we use chelating resins such as Chelex 100. For our most frequently used buffers, we make the buffer and store in large Erlenmeyer flasks (4-L). Some of the flasks have been in service for over a decade holding only a specific buffer. Thus, they transfer virtually no metal from the Pyrex glass to the solutions.

Use plastic ware to spoon out Chelex (chelating) resin so that no metal is introduced into the resin. Our usual approach is:

1. In general use 5 – 10 mL of resin per 1 liter of buffer.

2. Transfer resin to be used into a 50 mL conical bottom centrifuge tube (not polypropylene), and then

3. Add some buffer to wash the resin. (Resin is acidic and the wash will adjust the pH).

4. Centrifuge, gently, or let the resin settle and discard the supernatant. This not only adjusts pH but removes potential interfering substances from the resin [42].

5. Repeat steps 3 and 4 one more time.

6. Add washed resin to the batch of buffer that needs to be de-metalled.

7. Stir solution very slowly for approximately 24 h (resin needs time to chelate all the metals). Stir as slowly as possible so any moving and tumbling of the resin is nil or as minimal as possible. This is done to minimize mechanical breakage of the resin.

8. If buffer needs to be sterile, then filter through a 0.22 μm filter before use. The buffer can be stored over the resin for several months.

9. Clean resin has a bright white appearance. When the resin begins to show a rust-red (iron) or blue (copper) tinge, the resin should be renewed. The resin is expensive; it is easily recycled with acid washes to remove chelated metals, then pH adjustment.

2.10. Note on headspace air

Oxygen and catalytic metals are the enemies of aqueous solutions of ascorbate; vida supra. At room temperature, air contains 30 – 35 times more oxygen than an equal volume of water or buffer. Thus, minimization of the volume of air, i.e. oxygen, in the headspace can reduce the amount of ascorbate oxidized in stock solutions when stored for longer times.

2.11. Note on DHA and stability

Dehydroascorbic acid (DHA), the two-electron oxidation product of ascorbate/ascorbic acid readily hydrolyzes in aqueous solutions as a function of pH [17], Figs. 3 and 4. Velisek et al. demonstrated that DHA is most stable in the pH range of 3 – 4. At pH 5, the half-life of DHA is about 1800 min (30 h), which shortens to about 30 min at pH 7.4 and <1 min at pH 9.

Fig. 4.

The half-life (in minutes) of DHA in aqueous solution as a function of pH. The points were calculated from the data and analytical expressions provided by Velisek et al. [17]. These data should be used to guide the handling of DHA, but the exact half-life of DHA at a specific pH value will vary considerably, depending on the content of the solution.

However, it must be kept in mind that the composition of the buffer solution can greatly affect the rate of this decomposition. For example, DHA in Tris buffer and cell culture medium has a quite different, pH-dependent half-life [43]. The presence of bicarbonate accelerates the hydrolysis of DHA as it acts as a base; as an example, Koshiishi et al. found that in DMEM cell culture medium at pH 7.4 containing bicarbonate that t_1/2 < 2 min [43]. This contrasts with a half-life of about 90 min observed in standard phosphate buffered saline at room temperature [44].

Temperature is also a consideration for the rate of DHA hydrolysis; the higher the temperature the greater the rate of loss of DHA. This variable is especially important if samples are to be stored [44,45]. Taking into account these many variables, it is clear that the results of Velisek et al., Fig. 4, are valid for their specific experimental conditions. Thus, the data of Fig. 4 should be used to guide the handling of DHA, but the exact half-life at a specific pH value, temperature, and matrix may vary considerably.

The glutathione/glutaredoxin enzyme system (GSH/Grx) of cells is thought to be one of the major routes for recycling DHA back to AscH⁻ (K_M = 0.25 mM, V_max = 100 nmol (s•mg protein)⁻¹) [46]; this suggests that k_cat is about 1 s⁻¹ for human neutrophil Grxs. However, other enzyme systems such as thioredoxin reductase (TrxR) can also accomplish the recycling, (k_cat ≈ 1.5 s⁻¹) [47], similar to Grx. Thus, there are redundant enzyme systems for the recycling of DHA back to AscH⁻. The role played by each would be a function of the actual concentration of active sites for these enzymes, assuming efficient redox recycling of these enzymes. In red cells, the concentration of Grx is ≈ 6 μM while TrxR is only ≈ 0.4 μM (as active sites) [48]. Thus, the glutaredoxin system will be the primary route to reduction of DHA in RBCs. In HepG2 cells there is approximately twice as many copies of Grx compared to TrxR [49]. Thus, Grx will probably dominate the recycling with TrxR contributing a larger fraction of the recycling, compared to RBCs. Needless to say, each cell type will recycle DHA through uniquely different reductase systems and redox environments.

2.12. Note on exposure of cells in culture to ascorbate

When exposing cells we use only a 1-h exposure to high levels of ascorbate [35–37]. This produces a relatively constant flux of H₂O₂ over this timeframe. In the literature, others may introduce ascorbate but then do not remove it. We find this approach to yield less reproducible results, i.e., greater variation. The flux of H₂O₂ will not be constant, and the far downstream decomposition products of ascorbate, i.e. the breakdown products of dehydroascorbic acid (DHA) may have varying effects. This is avoided with the 1-h exposure in conjunction with specification of dose of ascorbate as mol cell⁻¹ [35]. Reporting ascorbate concentration, exposure time, media volume (including details about the medium), and cell numbers in methods sections are an aid to increased Rigor and Reproducibility.

2.13. Ascorbate phosphate vs. ascorbate

In cell culture experiments, the introduction of ascorbate will produce a flux of H₂O₂ in the culture medium and thereby pose an oxidative challenge to the cells [35,50]. This H₂O₂ could be the focus of a study if the pro-oxidant character of ascorbate is being examined. However, this oxidation would be an unwanted confounder if other aspects of ascorbate biochemistry are to be studied. To overcome this issue, esterified forms of ascorbate, such as l-ascorbic acid 2-phosphate, can be used that are resistant to oxidation and yet retain the activity of vitamin C [51–53]. To be active as vitamin C, the ester moiety must be removed; this can be accomplished via slow hydrolysis in the medium or most likely by a phosphatase enzyme [52]. The most widely used and readily available ester is l-ascorbic acid 2-phosphate. This is widely used as a source of vitamin C in aquaculture as it resists oxidation and has outstanding availability as vitamin C [54]. It has also been found to be of great use in cell culture due to its stability in aqueous environments. The accumulation of ascorbate in cells when l-ascorbic acid 2-phosphate is provided in the culture medium is somewhat slower than when ascorbate is the source [55], but confounding oxidation in the medium is avoided [56].

3.0. Results

3.1. Oxygen

When ascorbate oxidizes in aqueous solutions, it naturally passes electrons to dioxygen (O₂), thereby removing it from the system. The reaction typically is a 2-electron reduction of O₂ yielding H₂O₂, Rxn 1. (Fig. 1)

$${\text{AscH}}^{-}+{\text{O}}_2+{\text{H}}_2\text{O}+2{\text{H}}^{+}-\left({\text{Fe}}^{3+/2+}\right)\to {\text{H}}_2{\text{O}}_2+\text{DHA}\left({\text{C}}_6{\text{H}}_8{\text{O}}_7\right)$$ (1)

The H₂O₂ produced can be further reduced to water, with Rxn 2 being the overall reaction.

$${\text{AscH}}^{-}+{\text{H}}_2\text{O}+2{\text{H}}^{+}-\left({\text{Fe}}^{2+/3+}\right)\to {\text{H}}_2{\text{O}}_2+\text{DHA}$$ (2)

Fe^3+/2+ and Fe^2+/3+ represents the redox cycling of the catalytic metal as it facilitates the oxidation of ascorbate.

Note, Rxns 1 and 2 are balanced as one molecule of H₂O is taken up in the formation of DHA, as shown in Fig. 1. Also, Rxns 1 and 2 show overall reactions; the detailed mechanisms are varied and complex. However, it is long-known that these reactions are facilitated by redox active transition metals [57]. Although the Fe^3+/2+ redox pair is indicated above, other redox active metals can also catalyze the reactions. In the absence of these metals, these reactions are very, very slow, even in near-neutral and especially acidic solutions [3].

Although the production of H₂O₂ in Rxn 1 is stoichiometric, the amount of H₂O₂ that accumulates nearly always is less and can even be below the limit of detection due to Rxn 2. In our experience, if the level of transition metal catalyst is low, more H₂O₂ can accumulate; if the amount is high, then less H₂O₂ accumulates. The meaning of “low” and “high” is relative and depends on the actual composition of the system being studied.

The solubility of oxygen in water varies with temperature. At 5 °C the concentration of oxygen in air-saturated pure water will be about 400 μM [31]. The concentration of ascorbate in the IV bag was about 426 mM. Considering ionic strength due to the presence of ascorbate, the concentration of oxygen will be less than in pure water; estimated to be about 330 μM, i.e. 0.33 mM [31]. As the volume of solution in the bag is 1 L, the amount of oxygen in the solution will be about 330 μmoles = 0.33 mmoles. Note that the amount of ascorbate in the bag was 426 mM; thus, the ratio of ascorbate to oxygen in the solution phase was 426 mM/0.33 mM ≈ 1300.

The volume of air in the bag was about 100 mL, or 0.1 L. This is typical for bags being prepared for patient use in the clinic. Using the Ideal gas Law, PV = nRT, which rearranges to

$$n=\text{PV}/\text{RT}$$

$$n=(0.21\text{atm})(0.1\text{L})/\left(0.082{\text{L}\text{atm}\text{K}}^{-1}\cdot {\text{mol}}^{-1}\right)(278\text{K})$$

$$n=0.00092{\text{mole}\text{of}\text{O}}_2=0.92{\text{mmol}\text{of}\text{O}}_2\text{.}$$

Note that there is about 3x more O₂ in the 0.1 L of headspace air than in the 1 L of aqueous solution.

The total number of moles of O₂ in the bag will be 0.33 + 0.92 ≈ 1.25 mmol. Thus, the ratio of ascorbate to total oxygen is:

$$(\text{ascorbate})/(\text{total}\text{oxygen})=426\text{mmol}/1.25\text{mmol}=340$$

The total number of moles of ascorbate in the bag is about 426 mmol. It will require 1.25 mmol of this ascorbate to reduce this 1.25 mmol of O₂ to H₂O₂ and perhaps another 1.25 mmol if all of this H₂O₂ if reduced to H₂O. This will result in a loss of at most only ≈0.6% of the ascorbate. With the variation (noise) in the measurements of ascorbate using typical laboratory apparatus, e.g., adjustable pipettes, this change will not be detectable. Only if there is replenishment of the O₂ in the bag by diffusion of atmospheric O₂ through the bag, allowing for ongoing oxidation of ascorbate, will there be a measurable change in the concentration of ascorbate.

3.2. Loss of ASCH⁻

The actual rate of loss of ascorbate in the infusion bag is very, very slow; the slope being −0.178 mM h⁻¹, or 0.031 g h⁻¹, Fig. 5. Thus, over 24 h the loss will be about 4 mM or 0.75 g, about 1% of the total ascorbate present. This is about twice the amount predicted if all the oxygen in the bag were to be consumed and there was no diffusion of atmospheric oxygen into the bag. This suggests there is some diffusion of oxygen into the bag. This slow rate reflects the lack of catalytic metals as well as the slightly acidic pH (5.75) of the solution [3].

Fig. 5.

Change in the concentration or amount of ascorbate over time. The two panels show the same data with two different ordinates, concentration as mM, and amount in the bag as grams of ascorbate. The uncertainties presented represent standard errors. Visually this may appear as highly variable, but in reality, the variations are quite small and are within the error expected with the standard laboratory equipment and procedures employed. Note that the ordinate covers only 9% of the maximum.

The intercept from the linear regression analysis of the data match exceptionally well the predicted concentration (427 mM vs. predicted 426 mM) or amount of ascorbate (75.3 g vs. predicted 75.0 g) in the bag.

3.3. pH of solution

The pH of the solution of ascorbate was determined to be about 5.75. This value did not change over the duration of the monitoring. This indicates that the half-life of the DHA formed upon oxidation of ascorbate will be about 900 min (15 h); estimated from Fig. 4. Thus, only a small amount of DHA will be irreversibly lost. If some DHA is a part of the infusion, it will be efficiently taken up by red blood cells and rapidly reduced back to AscH⁻[44,58–61]; thus, only a small fraction of the ascorbate in the bag will actually be lost.

3.4. Buffer capacity

For this study, buffer capacity (β) is defined as the quantity of strong acid or base that must be added to change the pH of one liter of solution by one pH unit. As is common in the health sciences, we express buffer capacity in units of mM/(1 pH unit).

$$\beta =\Delta n/\Delta pH$$

where Δn is the number of moles of strong acid or base added to a 1 liter of a buffer solution and ΔpH is the resulting change in pH.

Using quantitative sequential additions of small, but equal amounts (moles) of acid or base (10 μL of 1.00 M) to a fixed volume of the ascorbate solution (8.0 mL), it is seen that the buffer capacity of the solution is essentially due to the ascorbate present, pK_a1 ≈ 4.0 at the ionic strength of the solution, Fig. 6. The change in pH upon addition of acid or base is not constant across the pH range examined. Larger pH changes are observed at higher pH values, as expected. This implies that the buffer capacity is not constant and decreases with increasing pH.

Fig. 6.

The pH of 8.0 mL of the solution of P-AscH⁻(75 g/L) upon addition of equal (moles) aliquots of acid or base. These data indicate that the acid-base properties of the ascorbic acid/ascorbate monoanion (AscH₂/ AscH⁻) equilibrium, pK_a1 (AscH₂ = 4.0) is the actual buffer, as expected. Other components of ASCOR^® are not significant contributors to the buffer capacity of the solution prepared for IV delivery to patients.

Using the data of Fig. 6, the buffer capacity of this solution at various pH values was determined, Fig. 7. The buffer capacity at around pH 6 is about 20 mM/pH. As expected, this drops precipitously as the pH increases so that at around pH 7.4 the buffer capacity is <2 mM/pH. The buffer capacity of blood is around 38 mM/pH [62]. Thus, infusion of P-AscH⁻appears to present little challenge to the pH of blood.

Fig. 7.

Buffer capacity of AscH⁻solution for IV delivery. The solution contains 75 g of ascorbate (150 mL ASCOR^®) in 850 mL of Sterile Water for Injection. The arrow marks the initial pH of the solution.

This is supported by the measurement of the pH values of plasma of samples collected from subjects immediately before and immediately after infusion of P-AscH⁻, Table 2. The similarity of pH values before and after infusion, suggest that the infusion produced no significant change in blood pH due to the high buffer capacity of blood and the low buffer capacity of P-AscH⁻at physiological pH.

Table 2.

Plasma pH values after processing^a.

Patient #	pH, pre-infusion	pH, post Infusion
A-01	7.71	7.71
A-01	7.61	7.62
A-02	7.53	7.61
G-05	7.97	8.06
G-05	7.78	7.72
G-06	7.64	7.60

^aBlood samples were collected in heparin Vacutainers. RBCs were removed to provide the plasma samples.

Although the data of Table 2 indicate no, or little change in pH upon infusion of P-AscH⁻, a weak point in this conclusion is that any buffer capacity due to the heparin vacutainer system used to collect the sample is not accounted for.

4.0. Conclusions

In this study we examined the stability of ascorbate in an IV bag as typically prepared by the pharmacy of the University of Iowa Hospitals and Clinics (UIHC) for use in the clinic. We found that:

1. The initial concentration of ascorbate in the bag to be 427 mM, matching the expected concentration of 426 mM (75 g/L).

2. We estimated that the available oxygen in the bag to be:

   1. 0.33 mmol in the aqueous phase;

   2. 0.92 mmol in the air-filled headspace, assuming its volume is about 10% of the solution volume;

   3. Total O₂ available = 1.25 mmol.

3. The ratio of (ascorbate) to (total oxygen) = 330; total O₂ is 0.3% of ascorbate.

4. The rate of ascorbate loss was about 0.04% h⁻¹. Over 24 h the loss was ≈4 mM or 0.75 g/L, about 1% of the total ascorbate present.

5. The buffer capacity of the ascorbate solution varies considerably between pH 6 (26 mM/pH) and 7.4 (≈2 mM/pH). But at pH 7.4, the buffer capacity is considerably less than blood, 38 mM/pH [62]; thus, little impact on blood pH would be expected upon infusion.

From the standpoint of the stability of ascorbate, preparing the IV bag, using current UIHC protocols, 24 h ahead of need appears not to be an issue.

In this work, we also share our experience on approaches to prepare and handle aqueous solutions of ascorbate to achieve the best Rigor and Reproducibility in all experiments focused on ascorbic acid/ascorbate, be they chemical, biochemical, cell culture, studies with animals, or for human clinical trials.

Data availability

Data will be made available on request.