Predicting Spray Dried Dispersion Particle Size Via Machine Learning Regression Methods.

Spray dried dispersion particle size is a critical quality attribute that impacts bioavailability and manufacturability of the spray drying process and final dosage form. Substantial experimentation has been required to relate formulation and process parameters to particle size with the results limited to a single active pharmaceutical ingredient (API). This is the first study that demonstrates prediction of particle size independent of API for a wide range of formulation and process parameters at pilot and commercial scale. Additionally we developed a strategy with formulation and target particle size as inputs to define a set of "first to try" process parameters. An ensemble machine learning model was created to predict dried particle size across pilot and production scale spray dryers, with prediction errors between -7.7% and 18.6% (25th/75th percentiles) for a hold-out evaluation set. Shapley additive explanations identified how changes in formulation and process parameters drove variations in model predictions of dried particle size and were found to be consistent with mechanistic understanding of the particle formation process. Additionally, an optimization strategy used the predictive model to determine initial estimates for process parameter values that best achieve a target particle size for a provided formulation. The optimization strategy was employed to estimate process parameters in the hold-out evaluation set and to illustrate selection of process parameters during scale-up. The results of this study illustrate how trained regression models can reduce the experimental effort required to create an in-silico design space for new molecules during early-stage process development and subsequent scale-up.

Feasibility of spectral pH measurement during the low-pH virus inactivation step of continuous therapeutic antibody production.

Acidic virus inactivation is commonly used during production of biotherapeutic products to provide virus safety in case of undetected virus contamination. Accurate pH measurement is required to ensure the product pH reaches a virus-inactivating level (typically 3.5-3.7), and a level post-inactivation that is appropriate for later purification steps (typically 5.5-7.5). During batch low-pH inactivation in discrete tanks, potentiometric glass probes are appropriate for measuring pH. During continuous inactivation for 2-3 weeks in an enclosed product stream, probe calibration drift and lag may lead to poor accuracy, and operational difficulties when compensating for drift. Monitoring the spectral response of compounds (indicators) in the product stream whose spectra are pH-sensitive offers a possible alternative way to measure pH without these drawbacks. Such indicators can already exist in the stream (intrinsic) or can be added (extrinsic). Herein are reported studies evaluating the feasibility of both.Promising ultraviolet screening results with the two extrinsics studied, thiamine and ascorbic acid, led to the addition of both to product stream samples titrated to different potentiometric pH values in the 3.3-4.5 range (a representative range encountered during continuous inactivation), and attempts to model pH using sample ultraviolet spectra. One model, based on variability in six spectral attributes, was able to predict pH of an independent sample set within ±0.07 units at the 95% confidence level. Since a typical inactivating pH tolerance is ±0.1 units, the results show that extrinsic indicators potentially can measure inactivation pH with sufficient accuracy. Suggested future steps and an alternative approach are presented.

CaM kinase control of AKT and LNCaP cell survival.

AKT and its substrate BAD have been shown to promote prostate cancer cell survival. Agonists, such as carbachol, and hormones that increase intracellular calcium concentration can activate AKT leading to cancer cell survival. The LNCaP prostate cancer cells express the carbachol-sensitive M(3) -subtype of G protein-coupled receptors that cause increases in intracellular calcium and activate the family of Ca(2+) /calmodulin-dependent protein kinases (CaM Ks). One type of CaM Kinase, CaM Kinase Kinase (CaM KK), phosphorylates several substrates including AKT on threonine 308. AKT phosphorylation and activation enhances cell survival through phosphorylation of BAD protein and the subsequent blockade of caspase activation. Our goals were to examine the mechanism of carbachol activation of AKT and BAD in LNCaP prostate cancer cells and evaluate whether CaM KK may be mediating carbachol's activation of AKT and cell survival. Our results suggest that carbachol treatment of LNCaP cells promoted cell survival through CaM KK and its phosphorylation of AKT. The bacterial toxin anisomycin triggered caspase-3 activation in LNCaP cells that was blocked by carbachol in a CaM KK- and AKT-dependent manner. AKT and BAD phosphorylation were blocked by the selective CaM KK inhibitor, STO-609, as well as siRNA directed against CaM KK. BAD phosphorylation was also blocked by treating cells with the AKT inhibitor, AKT-X, as well as siRNA to AKT. Additionally, epinephrine promoted LNCaP cell survival through activation of AKT that was insensitive to STO-609. Taken together these data suggest a survival role for CaM KK operating through AKT and BAD in LNCaP prostate cancer cells.

ERK activation and cell growth require CaM kinases in MCF-7 breast cancer cells.

Previous studies on MCF-7 breast cancer cells have shown that the G-protein coupled receptor (GPCR) agonist carbachol increases intracellular calcium levels and the activation of extracellular signal-regulated kinase (ERK). Calcium and calmodulin regulate the calcium/calmodulin-dependent kinase (CaM kinase) family of proteins that have been proposed to regulate ERK and gene transcription. Our results suggest that both estrogen (E2) and carbachol treatment of MCF-7 breast cancer cells trigger phosphorylation of ERK1/2 and the transcription factor Elk-1. Carbachol and estrogen triggered nearly a four- to sixfold increase in MCF-7 cell proliferation by 96 h, respectively. Carbachol-stimulated ERK activation and cell growth was completely blocked by the Muscarinic M(3)-subtype GPCR inhibitor, 4-DAMP, and siRNA against the M(3)-subtype GPCR. Interestingly, blockade of CaM KK with the selective inhibitor STO-609 prevented carbachol activation CaM KI, ERK, Elk-1, and cell growth. Consistent with these observations, knockdown of CaM KKalpha and CaM KIgamma with shRNA-containing plasmids blocked ERK activation by carbachol. In addition, Elk-1 phosphorylation and luciferase activity in response to carbachol treatment was also dependent upon CaM kinases and was inhibited by U0126, STO-609, and siRNA knockdown of CaM kinases and ERK2. Finally, blockade of either CaM KK (with STO-609) or ERK (with U0126) activities resulted in the inhibition of carbachol- and estrogen-mediated cyclin D1 expression and MCF-7 cell growth. Taken together, our results suggest that carbachol treatment of MCF-7 cells activates CaM KI, ERK, the transcription factor Elk-1, cyclin D1, and cell growth through CaM KK.

Calmodulin-dependent kinase kinase/calmodulin kinase I activity gates extracellular-regulated kinase-dependent long-term potentiation.

Intracellular Ca2+ and protein phosphorylation play pivotal roles in long-term potentiation (LTP), a cellular model of learning and memory. Ca2+ regulates multiple intracellular pathways, including the calmodulin-dependent kinases (CaMKs) and the ERKs (extracellular signal-regulated kinases), both of which are required for LTP. However, the mechanism by which Ca2+ activates ERK during LTP remains unknown. Here, we describe a requirement for the CaMK-kinase (CaMKK) pathway upstream of ERK in LTP induction. Both the pharmacological inhibitor of CaMKK, STO-609, and dominant-negative CaMKI (dnCaMKI), a downstream target of CaMKK, blocked neuronal NMDA receptor-dependent ERK activation. In contrast, an inhibitor of CaMKII and nuclear-localized dnCaMKIV had no effect on ERK activation. NMDA receptor-dependent LTP induction robustly activated CaMKI, the Ca2+-stimulated Ras activator Ras-GRF1 (Ras-guanyl-nucleotide releasing factor), and ERK. STO-609 blocked the activation of all three enzymes during LTP without affecting basal synaptic transmission, activation of CaMKII, or cAMP-dependent activation of ERK. LTP induction itself was suppressed 50% by STO-609 in a manner identical to the ERK inhibitor U0126: either inhibitor occluded the effect of the other, suggesting they are part of the same signaling pathway in LTP induction. STO-609 also suppressed regulatory phosphorylation of two downstream ERK targets during LTP, the general translation factors eIF4E (eukaryotic initiation factor 4) and its binding protein 4E-BP1 (eukaryotic initiation factor 4E-binding protein 1). These data indicate an essential role for CaMKK and CaMKI to link NMDA receptor-mediated Ca2+ elevation with ERK-dependent LTP.

Calcium activation of ERK mediated by calmodulin kinase I.

Elevated intracellular Ca(2+) triggers numerous signaling pathways including protein kinases such as the calmodulin-dependent kinases (CaMKs) and the extracellular signal-regulated kinases (ERKs). In the present study we examined Ca(2+)-dependent "cross-talk" between these two protein kinase families. Using a combination of pharmacological inhibitors and dominant-negative kinases (dnKinase), we identified a requirement for CaMKK acting through CaMKI in the stimulation of ERKs upon depolarization of the neuroblastoma cell line, NG108. Depolarization stimulated prolonged ERK and JNK activation that was blocked by the CaMKK inhibitor, STO-609; this inhibition of ERK activation by STO-609 was rescued by expression of a STO-609-insensitive mutant of CaMKK. However, activation of ERK by epidermal growth factor or carbachol were not suppressed by inhibition of CaMKK, indicating specificity for this "cross-talk." To identify the downstream target of CaMKK that mediated ERK activation upon depolarization, dnKinases were expressed. The dnCaMKI completely suppressed ERK2 activation whereas dnAKT/PKB or nuclear-targeted dnCaMKIV, other substrates for CaMKK, were not inhibitory. ERK activation upon depolarization or transfection with constitutively active (ca) CaMKI was blocked by dnRas. Additionally, depolarization of NG108 cells promoted neurite outgrowth, and this effect was blocked by inhibition of either CaMKK (STO-609) or ERK (UO126). Co-transfection with caCaMKK plus caCaMKI also stimulated neurite outgrowth that was blocked by inhibition of ERK (UO126). These data are the first to suggest that ERK activation and neurite outgrowth in response to depolarization are mediated by CaMKK activation of CaMKI.

Galpha and Gbeta gamma require distinct Src-dependent pathways to activate Rap1 and Ras.

The Src tyrosine kinase is necessary for activation of extracellular signal-regulated kinases (ERKs) by the beta-adrenergic receptor agonist, isoproterenol. In this study, we examined the role of Src in the stimulation of two small G proteins, Ras and Rap1, that have been implicated in isoproterenol's signaling to ERKs. We demonstrate that the activation of isoproterenol of both Rap1 and Ras requires Src. In HEK293 cells, isoproterenol activates Rap1, stimulates Rap1 association with B-Raf, and activates ERKs, all via PKA. In contrast, the activation by isoproterenol of Ras requires Gbetagamma subunits, is independent of PKA, and results in the phosphoinositol 3-kinase-dependent activation of AKT. Interestingly, beta-adrenergic stimulation of both Rap1 and ERKs, but not Ras and AKT, can be blocked by a Src mutant (SrcS17A) that is incapable of being phosphorylated and activated by PKA. Furthermore, a Src mutant (SrcS17D), which mimics PKA phosphorylation at serine 17, stimulates Rap1 activation, Rap1/B-Raf association, and ERK activation but does not stimulate Ras or AKT. These data suggest that Rap1 activation, but not that of Ras, is mediated through the direct phosphorylation of Src by PKA. We propose that the beta(2)-adrenergic receptor activates Src via two independent mechanisms to mediate distinct signaling pathways, one through Galpha(s) to Rap1 and ERKs and the other through Gbetagamma to Ras and AKT.

Crosstalk between cAMP and MAP kinase signaling in the regulation of cell proliferation.

Hormonal stimulation of cyclic adenosine monophosphate (cAMP) and the cAMP-dependent protein kinase PKA regulates cell growth by multiple mechanisms. A hallmark of cAMP is its ability to stimulate cell growth in many cell types while inhibiting cell growth in others. In this review, the cell type-specific effects of cAMP on the mitogen-activated protein (MAP) kinase (also called extracellular signal-regulated kinase, or ERK) cascade and cell proliferation are examined. Two basic themes are discussed. First, the capacity of cAMP for either positive or negative regulation of the ERK cascade accounts for many of the cell type-specific actions of cAMP on cell proliferation. Second, there are several specific mechanisms involved in the inhibition or activation of ERKs by cAMP. Emerging new data suggest that one of these mechanisms might involve the activation of the GTPase Rap1, which can activate or inhibit ERK signaling in a cell-specific manner.

PKA phosphorylation of Src mediates cAMP's inhibition of cell growth via Rap1.

In fibroblast cells, cAMP antagonizes growth factor activation of ERKs and cell growth via PKA and the small G protein Rap1. We demonstrate here that PKA's activation of Rap1 was mediated by the Rap1 guanine nucleotide exchange factor C3G, the adaptor Crk-L, the scaffold protein Cbl, and the tyrosine kinase Src. Src was required for cAMP activation of Rap1 and the inhibition of ERKs and cell growth. PKA activated Src both in vitro and in vivo by phosphorylating Src on serine 17 within its amino terminus. This phosphorylation was required for cAMP's activation of Src and Rap1, as well as cAMP's inhibition of ERKs and cell proliferation. This study identifies an antiproliferative role for Src in the physiological regulation of cell growth by cAMP.