The Impact of Instantaneous Ultra-High Temperature (INF) Versus Conventional Thermal Processing on Bovine Milk: Nutritional and Physicochemical Perspectives

TL;DR

This study compares different milk processing methods and finds that INF treatment preserves more nutrients and proteins while ensuring safety.

Contribution

The study introduces INF as a novel thermal processing method that better preserves milk's nutritional and immune-active components compared to UHT and conventional pasteurization.

Findings

INF treatment preserves native α-lactalbumin and β-lactoglobulin better than conventional methods.

INF retains higher levels of lactoferrin and immunoglobulin G compared to pasteurization and UHT.

INF results in lower glycation content and furosine concentration compared to conventional heat treatments.

Abstract

Balancing microbial safety and the retention of heat-sensitive components has long been a key issue in dairy processing research. This study systematically compared the effects of instantaneous ultra-high-temperature treatment (INF, 145–155 °C/0.09 s) with that of conventional pasteurization (75–95 °C/15 s) as well as ultra-high-temperature treatment (UHT, 135 °C/5 s), on the microbial evaluation, nutritional composition, and physicochemical quality of bovine milk. The results showed that all heat treatments completely inactivated Staphylococcus aureus, coliforms, while only UHT and INF achieved full spore elimination. In the INF group, α-lactalbumin remained almost completely native and native β-lactoglobulin retention was approximately 83% relative to raw milk. The retention of lactoferrin and immunoglobulin G was about 30% and 12% after INF treatment, respectively, which were higher than that of 13% and 8% in the 85 °C/15 s group, and complete denaturation in the 95 °C/15 s and UHT groups. Furthermore, vitamin B2 remained stable after INF treatment. The glycation content of proteins was lower in INF treatment compared to conventional heat treatments, especially for the concentration of furosine, which was about 6–7 mg/100 g protein in the INF group, and 15 mg/100 g protein in the 95 °C/15 s group, and 67 mg/100 g protein in the UHT group. Overall, the INF process achieved sterilization equivalent to UHT while substantially reducing thermal load, thus better balancing microbial safety, nutritional integrity, and immune-active proteins, which provides a scientific basis for establishing standardized INF parameters and promoting high-quality dairy production.

Figures

Full text

1. Introduction

Bovine milk is a major natural source of high-quality proteins, lipids, vitamins, minerals, and bioactive compounds in the human diet. Conventional thermal treatments such as pasteurization (HTST, 72–75 °C/15 s) and ultra-high-temperature processing (UHT, 135 °C/5 s) are widely used in the dairy industry, but they present a clear trade-off. HTST pasteurization better preserves whey proteins and bioactive components, yet provides limited control over heat-resistant spores and results in a relatively short shelf life. In contrast, UHT can achieve commercial sterility, but leads to extensive whey protein denaturation and complete inactivation of lactoferrin and immunoglobulins, accompanied by nutritional losses and the development of “cooked” off-flavors [1,2], thereby compromising the nutritional value and functional properties of the product [3]. Consequently, achieving an optimal balance between microbial safety and nutrient preservation has long been a central challenge in dairy processing research.

Beyond heat treatment of liquid milk, more intensive thermal conditions are commonly applied in the production of infant formula. In industrial practice, a high-temperature pasteurization step at around 90–95 °C for 15 s (or similar time–temperature combinations) is usually introduced before evaporation and spray drying to sterilize the liquid formula base [4], ensuring microbial safety, inactivating heat-stable enzymes, and improving the storage stability of the final powder. However, compared with conventional HTST, such conditions substantially increase the thermal load, exacerbating the denaturation and aggregation of whey proteins, promoting lactose–protein glycation reactions, and further damaging heat-sensitive bioactive proteins such as LF and IgG. Therefore, there is an urgent need for alternative heat treatment strategies that can simultaneously meet the stringent safety requirements of infant formula while minimizing the loss of nutritional and functional components.

INF processing, characterized by “high temperature–ultra-short time”, has recently gained attention for its ability to achieve UHT-content sterilization with lower heat load. INF rapidly heats milk via direct steam injection or infusion, followed by instantaneous cooling, minimizing protein denaturation and Maillard reaction products (e.g., furosine). Studies have shown that INF-treated milk retains higher content of α-lactalbumin, β-lactoglobulin, LF, and IgG, with sensory and nutritional qualities closer to pasteurized milk [1,5]. Despite these advantages, the absence of standardized INF parameters and clear sterilization equivalence to HTST or UHT limits its industrial adoption.

In milk, bioactive proteins such as LF and IgG are crucial indicators of thermal impact due to their high heat sensitivity, while having immunological and antibacterial activities. Similarly, β-Lg the dominant whey proteins, play essential roles in heat-induced aggregation and milk stability. Their denaturation, interactions with caseins, and adsorption onto the milk fat globule membrane (MFGM) influence both microstructure and functional properties [6,7]. Other heat sensitive indicators, including alkaline phosphatase (ALP), lipase, and vitamins, also reflect the effect of heat load on product quality.

This study performed a systematic comparison of the physicochemical and nutritional quality of milk from pasteurization (75–95 °C/15 s) to conventional UHT (135 °C/5 s) to INF (145–155 °C/0.09 s) by analysis microbial, nutritional, and protein analyses—including LF, IgG, vitamin stability, and protein glycation. The findings of this study provide a scientific basis for identifying conventional heat treatment conditions that are equivalent to INF, clarifying its advantages over conventional sterilization, thereby supporting the determination of its process parameters and industrial application standards.

2. Materials and Methods

2.1. Sample Collection and Heat Treatments

Raw bovine milk was collected from Zhengyang Dairy Farm (Leyuan Animal Husbandry, Shijiazhuang, Hebei, China). After microbial sampling, the milk was frozen at −80 °C as the unheated control. Under laboratory conditions, the milk was subjected to seven thermal treatments: raw (control); HTST pasteurization at 75 °C/15 s; high temperature pasteurization at 85 °C/15 s and 95 °C/15 s; conventional UHT at 135 °C/5 s; and INF processing at 145 °C/0.09 s and 155 °C/0.09 s using an INF system developed by Junlebao Dairy. Heated samples were aseptically collected, rapidly cooled below 10 °C for microbiological tests, and the remaining aliquots stored at –80 °C for all other analyses. Milk was collected from cows in mid-lactation (i.e., established lactation), and reagents were of analytical grade (All samples were subjected to standardized fat-content adjustment and homogenization).

2.2. Microbiological Analysis

Microbiological quality was assessed by determining the total aerobic bacterial count, coliforms, S. aureus, spores, and psychrotrophic bacteria [8,9]. Total aerobic counts were determined using 3M Petrifilm™ Aerobic Count plates (Cat. No. 6406; Neogen, Lansing, MI, USA), while coliforms and S. aureus were enumerated using 3M Petrifilm™ Coliform Count (Cat. No. 6416; Neogen, Lansing, MI, USA) and Staph Express Count plates (Cat. No. 6491,6493; Neogen, Lansing, MI, USA), respectively. Spore counts were measured following the method of Moraru et al. [10].

2.3. Stability of Heat-Labile Components

2.3.1. Determination of Native Whey Proteins

Milk samples were centrifuged at 3000× g for 15 min at 4 °C to remove fat. The skimmed milk was adjusted to pH 4.6 with 1 mol/L HCl to precipitate caseins and denatured whey proteins, then equilibrated at 4 °C for 30 min. After confirming pH stability, the samples were ultracentrifuged at 100,000× g for 90 min at 25 °C, and the supernatant was collected as native whey [11]. Native whey protein content was determined by the Kjeldahl method and expressed as the percentage retained relative to raw milk.

2.3.2. SDS-PAGE of Native Whey Protein

SDS–PAGE was conducted according to Schägger [12] using a 12% separating gel and a 4% stacking gel. Whey samples (2 mg/mL) were mixed 1:1 (v/v) with 2× loading buffer; for reducing conditions, 5% β-mercaptoethanol was added. Samples were boiled for 3 min, cooled, and 10 μL was loaded per lane. Electrophoresis was performed at 60 V until the samples entered the separating gel, then at 120 V until completion. Gels were stained with Coomassie Brilliant Blue R-250 and destained until clear. Images were captured using a ChemiDoc XRS+ system (Bio-Rad, Hercules, CA, USA). Each sample was analyzed in duplicate.

2.3.3. Determination of Lactoferrin (LF) and Immunoglobulin G (IgG) by ELISA

LF and IgG in the native whey were quantified using commercial ELISA kits (Bethyl Laboratories, Montgomery, TX, USA) (Cat. No. E11-126, 118) following Liu et al. [11] and Heidebrecht et al. [13]. Whey samples were diluted 100–10,000 fold within the standard curve range, and 100 μL of each diluted sample or standard was added to antibody-coated 96-well plates. After incubation, washing, and enzyme-conjugate reactions, color was developed with TMB substrate and stopped with sulfuric acid. Absorbance was measured at 450 nm, and concentrations of LF and IgG were calculated by four-parameter logistic fitting (SoftMax Pro, Molecular Devices, San Jose, CA, USA) and expressed as retention percentages relative to raw milk.

2.3.4. Determination of α-Lactalbumin (α-La) and β-Lactoglobulin (β-Lg)

Whey samples were analyzed by reversed-phase HPLC using a C4 column (250 mm × 4.6 mm, 3.5 μm) (Alliance e2695 Separations Module; Waters, Milford, MA, USA). The mobile phases were 0.1% trifluoroacetic acid in water (A) and in acetonitrile (B) under gradient elution. The flow rate was 1.5 mL/min, the column temperature 60 °C, the detection wavelength 210 nm, and the injection volume 30 μL. α-La and β-Lg were identified by retention time matching with standards and quantified using external calibration. Concentrations of β-Lg A and β-Lg B were combined. Each sample was analyzed in duplicate, with relative standard deviation below 10%.

2.3.5. Determination of Vitamin C and Vitamin B2

Vitamins C (VC) and B_2_ (VB_2_) were quantified by ultra-high-performance liquid chromatography (UHPLC) following Tao et al. [14]. Milk samples were ultrasonicated under dark conditions, mixed with methanol (1:4, v/v), vortexed, and centrifuged at 10,000× g for 10 min at 4 °C. The supernatant was evaporated under nitrogen, reconstituted in ultrapure water/ether (1:1, v/v), and filtered through a 0.22 µm membrane. UHPLC analysis was performed on an HSS T3 column (100 mm × 2.1 mm, 1.8 µm, Alliance e2695 Separations Module; Waters, USA) at 40 °C with a 10 µL injection volume. The mobile phases were solvent A (0.01% formic acid and 10 mmol/L ammonium formate in water) and solvent B (5 mmol/L ammonium formate in methanol).

2.4. Microstructural Effects: Confocal Laser Scanning Microscopy (CLSM) Analysis

CLSM (LSM 880; Carl Zeiss, Oberkochen, Germany) was performed according to the method of Pu et al. [15] with minor modifications. Nile Red (1 mg/mL in ethanol) was prepared as a lipid-specific fluorescent dye, and Fast Green FCF (1 mg/mL in ultrapure water) was prepared as a protein-specific dye. For dual staining, 100 μL of each dye solution was added to 1 mL of milk sample, followed by incubation at 4 °C in the dark for 30 min. Approximately 10 μL of the stained emulsion was then placed on a glass slide, covered with a coverslip, and observed under CLSM. All dye solutions were stored at low temperature under dark conditions until use.

2.5. Determination of Carbonyl and Sulfhydryl Content

Protein carbonyl content was determined according to Levine et al. [16]. A UV–Vis spectrophotometer (UV-2700, Shimadzu, Kyoto, Japan) equipped with quartz cuvettes was used for absorbance measurement. Briefly, 0.5 mL of milk was mixed with 0.5 mL of DNPH solution (10 mol/L, prepared in 2 mol/L HCl) and incubated at room temperature for 1 h. Subsequently, 1 mL of 20% (w/v) trichloroacetic acid (TCA) was added, and the mixture was centrifuged at 10,000× g for 5 min. The supernatant was discarded, and the pellet was washed three times with an ethyl acetate/ethanol solution (1:1, v/v) to remove unreacted DNPH. Finally, the precipitate was dissolved in 0.5 mL of 6 mol/L guanidine hydrochloride solution. The blank was prepared by replacing milk with 2 M HCl and processed identically to the samples. Absorbance was recorded at 370 nm. Carbonyl content was calculated using a molar extinction coefficient of 22,000 L·mol^−1^·cm^−1^ and expressed as nmol/mg protein.

Protein thiol (sulfhydryl) content was measured according to Beveridge et al. [17] Briefly, 0.5 mL milk was mixed with 4.0 mL Tris–glycine buffer, followed by addition of 0.5 mL DTNB solution (10 M), and incubated for 30 min. The blank contained Tris–glycine buffer and DTNB reagent and was treated identically to the samples. Absorbance was measured at 412 nm. Thiol content was calculated using a molar extinction coefficient of 13,600 L·mol^−1^·cm^−1^ and expressed as nmol/mg protein.

2.6. Determination of Component Interactions

2.6.1. The Interactions Between Carbohydrate and Milk Proteins: Glycation Content of Whey Protein

The glycation content of whey proteins was analyzed by ultra-performance liquid chromatography coupled with electrospray ionization mass spectrometry (UPLC–ESI–MS, LCZ/2690 XE/996; Waters, Milford, MA, USA) [18] Whey samples (1 mg/mL) were separated on a BEH C4 column (2.1 × 100 mm, 1.7 μm) at 0.3 mL/min using a gradient of 0.1% formic acid in water (A) and acetonitrile (B): 0–8 min, 98% → 60% A; 8–10 min, 60% → 20% A. Mass spectrometry was performed on a SYNAPT Q-TOF instrument in positive ion mode (cone voltage 30 V, collision energy 6 eV, m/z 20–2000). Data were processed with MassLynx V4.1 software.

2.6.2. The Interactions Between Carbohydrate and Milk Proteins: Furosine Content

Furosine content was quantified according to D. Martysiak [19] and Guo et al. [20] with slight modifications. Milk samples (2 mL) were hydrolyzed with 6 mL of 10.6 mol/L HCl at 110 °C for 22 h. After cooling and filtration, the filtrate was analyzed by HPLC. Separation was performed on a C18 column (250 × 4.6 mm, 5 μm; Agilent Technologies, Santa Clara, CA, USA) at 32 °C, with detection at 280 nm. The mobile phases were 0.1% trifluoroacetic acid in water (A) and methanol (B), using a gradient of 0–13.2% B (0–16 min), 13.2–100% B (16–16.5 min), and 100–0% B (16.5–25 min). The flow rate was 0.5 mL/min, and furosine content was expressed as mg/100 g protein.

2.7. Statistical Analysis

Data were analyzed by one-way analysis of variance (ANOVA) and Duncan’s test using SPSS 21.0 software and expressed as mean ± standard deviation. Differences between samples were statistically significant at p < 0.05.

3. Results

3.1. Microbial Evaluation

The total viable count, coliforms, spores and S. aureus were analyzed in raw and heat-treated milk under various processing conditions. As summarized in Figure 1, raw milk contained 3.83 log CFU/mL total bacteria, with coliforms and spores at 2.76 and 2.07 log CFU/mL, respectively, and S. aureus at 2.13 log CFU/mL. Thermal processing markedly reduced microbial loads, and the extent of microbial inactivation increased with rising temperature. After HTST pasteurization at 75 °C and 85 °C, total viable counts decreased to 2.32 and 2.28 log CFU/mL, respectively—approximately 1.5 log reductions compared with raw milk—while coliforms and S. aureus became undetectable and only small numbers of spores remained (1.59 and 1.55 log CFU/mL). Treatment at 95 °C further reduced the viable and spore counts to 2.15 and 1.38 log CFU/mL, indicating partial spore inactivation. Complete microbial elimination was achieved under both UHT and INF treatment.

3.2. Preservation of Heat-Sensitive Components

3.2.1. Retention of Major Whey Proteins

To investigate the thermal stability of whey proteins in bovine milk subjected to different heat treatments, caseins and denatured whey proteins were separated from each sample, and the concentrations of native whey proteins were determined. As shown in Figure 2A, the native whey protein content in the 75 °C group was retained at approximately 95% compared to raw milk, whereas it was about 80% in the INF group, which was higher than that in the 85–95 °C groups (approximately 72% at 85 °C and 52% at 95 °C) and the UHT group (with a retention of around 40%).

As illustrated in Figure 2B, the retention patterns of α-La and β-Lg also differed significantly across treatments. Both proteins exhibited a gradual decrease in retention from 75 °C to 95 °C and reached the lowest contents after UHT processing. In the INF 155 °C groups, the denaturation contents of α-La and β-Lg (approximately 3% and 18%, respectively) were slightly higher than those in HTST milk at 75 °C, but were markedly lower than those observed at 95 °C (around 30% and 80%) and in UHT (around 60% and 90%). In contrast, in the INF 145 °C group, the α-La retention was almost comparable to that of raw milk, and although slightly reduced at 155 °C it remained higher than in the 85 °C, 95 °C and UHT groups.

Quantitative analysis of heat-sensitive functional proteins (Figure 2C) revealed distinct thermal susceptibilities between LF and IgG. Both proteins showed the highest concentrations in raw milk, followed by a sharp decline after treatment at 75 °C (LF: 65%; IgG: 75%). The retention of LF and IgG decreased to 13% and 8%, respectively, at 85 °C and were undetectable at 95 °C and UHT conditions. In contrast, the retention of LF and IgG was approximately 30% and 12%, respectively, in the INF treatments at both 145 °C and 155 °C, which was significantly higher than in the 85–95 °C and UHT groups.

SDS–PAGE results (Figure 2D) further supported these findings. The LF and IgG bands became faint after heating at 85 °C and 95 °C. The bands of α-La and β-Lg were also markedly weakened or blurred. In contrast, the INF-treated samples displayed much darker bands for LF, IgG, and major whey proteins than those observed in the 85 °C, 95 °C, and UHT groups, indicating greater preservation of native conformations and less protein aggregation.

3.2.2. Heat Stability of Vitamins

As shown in Figure 3, different heat treatments had minimal effects on VB_2_ content in milk. The VB_2_ content in raw milk was approximately 0.11 mg/100 g, and only the UHT treatment caused a slight but significant decrease in VB_2_ content to about 0.10 mg/100 g (p < 0.05), while the VB_2_ content in the other groups remained similar to that of raw milk. VC was not detected in any sample, likely due to its low natural concentration and high thermal and oxidative instability during processing.

3.3. Microstructural Changes in Fat–Protein Distribution Observed by Confocal Laser Scanning Microscopy

As shown in Figure 4, CLSM revealed that different heat treatments distinctly affected the spatial distribution and stability of fat droplets and proteins in bovine milk. In raw milk, fat exists as globules surrounded by a natural milk fat globule membrane (MFGM) mainly composed of lipoproteins. The fat globule diameters varied widely but were substantially reduced after homogenization (Figure 4A).

UHT treatment caused extensive aggregation between fat droplets and proteins, resulting in large fused complexes that became more pronounced after 14 days of storage. In contrast, HTST-treated samples exhibited only minor changes, maintaining a relatively uniform distribution. Remarkably, milk processed by INF treatments at 145 °C and 155 °C retained homogeneously dispersed fat and protein structures at both day 0 and day 14, with almost no visible aggregation.

3.4. Molecular Modifications of Milk Proteins: Glycation (Maillard Reaction) and Oxidation

Protein glycation and Maillard reaction markers were evaluated to understand the extent of protein-sugar interactions under different thermal treatments. Figure 5A,B show the UPLC-MS deconvoluted mass spectra of α-La and β-Lg, respectively. In raw milk (R), the dominant peaks for both proteins were located at approximately 14,177 Da for α-La and 18,362/18,363 Da for β-Lg, corresponding to their non-glycated monomers. After treatment at 95 °C and UHT, additional peaks (+Lac_1_, ~324 Da above the main peaks) were observed for both proteins, indicating the formation of glycation products due to lactose addition. However, no well-defined +Lac_2_ peaks (+648 Da) were detected, with only faint shoulder signals appearing in a few samples. β-Lg exhibited more prominent glycation compared to α-La. In contrast, after treatments at 75 °C, 85 °C, and INF (145–155 °C), both proteins showed spectra similar to that of raw milk, with the non-glycated monomer peaks dominating and minimal +Lac_1_ signals.

Furosine, a well-established marker for the early Maillard reaction between lactose and free amino groups in proteins, was also quantified (Figure 5C). The furosine content in raw milk was approximately 6.23 mg/100 g protein. Following 75 °C and INF (145–155 °C) treatments, furosine content did not significantly differ from raw milk. However, with increased heat intensity, particularly at 85 °C, 95 °C, and UHT, furosine content increased significantly; at 95 °C, furosine content doubled, and UHT resulted in the highest content of approximately 67.21 mg/100 g protein.

Heat treatment can induce the formation of protein oxidation products. Figure 6 presents the protein oxidation results, expressed as carbonyl and sulfhydryl contents. Compared with the control (R), heat treatment generally increased protein carbonyl contents while decreasing sulfhydryl contents. Specifically, the sulfhydryl content in the UHT 135 °C/5 s group decreased significantly from 6.77 to 3.04 nmol/mg protein, and that in the 95 °C/5 s group decreased from 6.77 to 4.90 nmol/mg protein. The sulfhydryl content in the 85 °C/15 s and INF-treated groups also showed a certain degree of reduction (approximately 1 nmol/mg protein). In addition, the protein carbonyl content in the UHT 135 °C/5 s group increased from 0.93 to 3.11 nmol/mg protein after heat treatment, whereas only a slight increase was observed in the INF group. These results indicate that protein oxidation occurred in all heat-treated samples; however, the extent of oxidation in the INF group was lower than that in the UHT 135 °C/5 s and 95 °C/15 s treatments.

4. Discussion

This study focused on instantaneous ultra-high-temperature processing (INF, 145–155 °C/0.09 s) and systematically compared it with traditional pasteurization (75–95 °C/15 s) and conventional ultra-high-temperature sterilization (UHT, 135 °C/5 s) in terms of microbial safety, heat-sensitive nutrients, and structural properties. Under microbial conditions where INF exceeded all pasteurization treatments including 95 °C treatment, INF markedly reduced the thermal loss of heat-labile components such as whey proteins, immune-active proteins, while also attenuating lactose-protein glycation and protein oxidation products. This trend is consistent with previous reports on direct steam infusion treatment: studies by Lu [21] and others have shown that INF processing can meet microbial inactivation and shelf-life requirements, while significantly lowering the contents of heat-load indicators, including furosine (10 mg/100 g protein) [21] and preserving more native whey proteins and functional components. The present findings provide a basis for benchmarking INF against conventional thermal processing by identifying temperature–time combinations (within the tested range) that yield comparable overall thermal intensity, as reflected by microbial inactivation, retention of heat-sensitive components, and protein glycation markers. As a result, the overall quality attributes of INF-treated milk more closely resembled those of medium- to low-intensity pasteurization rather than conventional UHT.

From the perspective of microbial evaluation, INF treatment in this study achieved complete elimination of the total viable count and spores (Figure 1). Pasteurization at 75–95 °C effectively inactivated coliforms and S. aureus, but showed limited lethality against heat-resistant spores (Figure 1), which is consistent with the findings of Wang [22] and Liu [23]. This indicates that although pasteurization can meet safety requirements, it cannot achieve full commercial sterility [24,25]. In addition, after treatment at 95 °C/15 s, a small number of viable cells (1.38 log CFU/mL) were still detected, whereas both UHT and INF treatments resulted in complete microbial inactivation. Similarly, Lu [21] reported that treatment at 135–147 °C/0.09 s achieved total microbial inactivation, and Wang [24] also observed complete microbial elimination using INF processing. These results collectively demonstrate that, INF is superior to any pasteurization treatment.

From the perspective of nutrient and functional component retention, INF generally was comparable to 75–85 °C/15 s for most indicators. In terms of native whey protein denaturation, the denaturation rates of α-La and β-Lg in the INF 155 °C/0.09 s group were approximately 3% and 18%, respectively. In Wang’s study [24], the denaturation contents of α-La and β-Lg in milk treated by INF (122–137 °C/0.116–3 s) were about 15% and 35%. Similarly, Mi [26] reported that in direct steam injection milk, the mean content of native α-La did not differ significantly from that in raw milk, while the mean content of native β-Lg decreased by about 30%, with no significant differences among treatments at 154, 155, and 156 °C for 0.116 s. These findings indicate that the trends observed in the present study are generally consistent with previous reports. Secondly, the denaturation rates of α-La and β-Lg in the INF group lay between those observed at 75 °C (not significantly different from raw milk) and 85 °C (approximately 6% and 34%, respectively) (Figure 2B), and were much lower than those in the UHT group (approximately 60% and 92%, respectively). In Wang’s study [24], the denaturation of α-La and β-Lg in INF-treated milk was also only slightly higher than in pasteurized milk (72 °C/15 s, about 12% and 20%), and far lower than in UHT-treated milk at 137 °C/4 s (approximately 73% and 88%, respectively). Other researchers comparing β-Lg denaturation in pasteurized and UHT milk have likewise reported that pasteurization generally causes only about 20% denaturation, whereas UHT treatment can lead to almost complete denaturation [27]. These results indicate that INF achieves a whey protein retention rate comparable to that of pasteurization.

The high whey protein retention in the INF-treated group is closely related to the unique working principle of INF: it heats the product by either directly injecting superheated steam in-line into the product (direct steam injection) or by passing the product through a steam-filled chamber (direct steam infusion). Compared with indirect tubular or plate heat exchangers, INF enables rapid and efficient heat transfer, thereby reducing unnecessary heat exposure and associated quality losses. Because it lacks heated surfaces, INF also minimizes burn-on and fouling. Consequently, INF is considered less detrimental to liquid proteins than indirect heating and has been increasingly used to produce extended shelf-life dairy products while limiting whey protein denaturation [28,29]. Therefore, although INF operates at a higher processing temperature, the actual exposure time is extremely short. Under such ultra-short heating conditions, even at a high peak temperature, a larger proportion of non-denatured whey proteins can still be retained in INF-treated samples.

In the immune protein, the INF group falls between 75 °C and 85 °C and is markedly superior to 95 °C and UHT. After INF treatment, LF and IgG were still retained for approximately 30% and 12%, respectively, which was significantly higher than at 85 °C (LF 13% and IgG 8%), whereas both LF and IgG were completely inactivated in the 95 °C and UHT groups (Figure 2C). In this study, the retention of LF after INF treatment was slightly lower than that reported previously. In Wang’s study [24], the retention of lactoferrin in INF-treated samples was about 45%, while Mi [26] reported that the denaturation rate of lactoferrin in milk samples treated with directly direct steam injection at 153.86 °C for 0.116 s was about 60%, both higher than the values observed in the present work. This discrepancy may be partly attributable to the lower processing temperatures used by Wang (122–137 °C/0.116–3 s) and partly to differences in milk source, as this study used bovine milk whereas Mi used yak milk. Nevertheless, our results still demonstrate that, in terms of immune proteins, the INF group lies between 75 °C and 85 °C/15 s and is clearly superior to 95 °C/15 s and UHT.

During milk processing, intermolecular interactions are mainly manifested as Maillard reactions between reducing sugars and proteins. Specifically, the early step involves the reaction of lactose with amino groups on basic amino acids (e.g., lysine, arginine) in proteins to form Amadori products. These products are commonly characterized by measuring their derivative, furosine [30]. Therefore, determining furosine content is of importance for evaluating the intensity of heat treatment in dairy products, quality control, and the retention of nutrients of infant formula. With respect to glycation indicators, INF more closely resembles HTST. Both the extent of lactose modification of α-La and β-Lg (Figure 5A,B) and the quantitative results for furosine (Figure 5C) indicate that INF treatment induces a relatively low content of Maillard reaction. This finding is consistent with the results of Lee [31], who reported that the average furosine content in indirectly heated UHT (IND-UP) milk was much higher than that in directly steam-injected UHT (DSI-UP) milk, and also aligns with the trend observed by Wang [24]. Taken together, these results suggest that, in terms of early Maillard reaction, INF is essentially equivalent to 75 °C/15 s conditions and clearly superior to 95 °C/15 s and UHT.

During thermal processing, in addition to interactions between proteins and lactose, interactions also occur between protein molecules themselves, such as those between whey proteins and/or caseins and the milk fat globule membrane. CLSM electrophoresis analyses (Figure 4) showed that under UHT conditions, extensive high-molecular-weight aggregates were formed and severe fat-protein flocculation occurred, whereas in the INF group, microscopic observations revealed a more uniform distribution of fat and protein and better overall physical stability compared with UHT. Although the processing temperature of the INF (instantaneous ultra-high-temperature) treatment is higher than that of UHT, its holding time is markedly shorter, so the actual thermal load experienced by proteins during processing is in fact lower than under UHT conditions. This finding indicates that thermal load is determined not by temperature alone, but by the combined effect of temperature and time. A lower thermal load helps to mitigate protein denaturation and progression of the Maillard reaction, which is particularly important for preserving the nutritional quality of dairy products and controlling furosine formation.

Against this background, integrating the above indicators, the INF process demonstrates a unique potential to balance sterilization efficiency and nutrient retention. In liquid milk processing, INF combines microbial inactivation comparable to UHT with thermal damage to milk components similar to pasteurization (within the 75–85 °C/15 s range), thereby enabling better retention of proteins, bioactive components, and reduce the degree of Maillard reaction. This suggests that INF could serve as a feasible technological option for developing higher-quality extended shelf-life (ESL) dairy products. In infant formula production, INF has the potential to serve as an alternative for intense pasteurization (95 °C/15 s). By ensuring while improving the retention of immune-active proteins such as LF and IgG and reducing the extent of Maillard reactions, INF provides a new strategy for reducing reliance on external fortification and for optimizing the nutritional and functional properties of formula compositions.

5. Conclusions

This study, through a systematic comparison of indicators, is the first to quantitatively propose the “equivalent thermal intensity range” of the INF process in key quality dimensions, providing empirical evidence for its potential applications in liquid milk. The findings not only offer data support for optimizing conventional thermal processing technologies, it also provides an innovative processing route for dairy product development. However, this study also has certain limitations. The current data do not yet cover important practical factors such as flavor compounds, sensory attributes, long-term shelf-life stability, and the cumulative thermal load arising from multiple heating steps in real industrial production. The conclusions are primarily based on a single-step thermal treatment model under laboratory conditions. Therefore, future work should systematically validate and refine the overall effects of INF under full-scale production conditions, integrating flavoromics, proteomics, and long-term storage trials, in order to further study the effective application of INF technology in industrial practice.