Written by Jan Majer (ESR13)

Before a novel medicine can be introduced on the market, it needs to undergo a drug development process which today on average consists of six stages – drug discovery, pre-clinical testing, phase I, phase II and phase III clinical testing and finally drug licensing. The drug development process from target discovery to the launch can take up to 15 years [1]. Moreover, the vast majority of pharmaceutical candidates fail in the pipeline causing so called “attrition”. The pressure of attrition drives scientists to discontinue research of candidates in very early stages of drug development during pre-clinical studies. Indeed, high attrition rates are steadily decreasing in all clinical stages over the recent years [2]. The application of coherent Raman scattering (CRS) has been one of many drivers behind the attrition rate decrease and continues to take up role in various pharmaceutical pipeline stages. CRS has drawn substantial attention since the beginning of the millennium spanning over plethora of biologically relevant applications including pharmaceutical tablet mapping, histology, drug delivery, ion transport and lipid disease studies.

Tablet formulation of a drug is commonly associated with Raman in pharmaceutical research. Drug tablets generally consist of an active pharmaceutical ingredient (API) – the active drug substance itself – and also “excipients” (e.g. polyethylene glycol, corn starch, and mannitol etc.) which stabilize and support the API. CRS has been widely implemented in understanding the crystallisation polymorphism which is known to impact solubility and bioavailability of an API. In Raman spectra, polymorphism can manifest in spectral shifts on a scale of tens of wavenumbers. Such shifts are distinctive enough in stimulated Raman scattering (SRS) -generated spectra to allow highly precise analysis of drug tablets as shown in figure 1 [3].

Furthermore, SRS is increasingly utilized in analysis of biological samples and has proven highly efficient in histology research. In histology, haematoxylin & eosin (H&E) staining has been a gold standard for decades owing to its robustness in cells and nuclei detection and allowing fast imaging and analysis in brightfield optical microscopy. However, a pitfall of H&E is the sample preparation. Despite the highly streamlined procedures today decreasing the preparation time of H&E samples to ~20-30 min, stimulated Raman histology (SRH) offers significantly faster alternative in terms of sample preparation while achieving the same tissue and cell segmentation data (figure 2) without any sample treatment, fixing or staining [4]. Moreover, in combination with deep learning methods SRH has been employed in brain tumour analysis with highly precise distinction rate of tumour and non-tumour tissues while using only two Raman shifts – 2850 cm-1 and 2950 cm-1 attributed to CH2 and CH3 molecular stretches, respectively. Due to the significantly shorter time period required for data extraction, SRH can be utilized in interoperational screening allowing the surgical team to quickly evaluate the tissue morphology and decide the direction of the procedure [5]. Moreover, CRS optical imaging can be enhanced by incorporating further optical modalities such as second harmonic generation (SHG). This is exploited for imaging of non-centrosymmetric structures, which in biological environment are typically collagen fibres and muscle fibres. Sarri et. al. lead a comparative study between the combined SRH and SHG with H&E staining of cryogenic biopsies of healthy and cancerous colon. Herein, Sarri simultaneously targets CH2 and CH3 stretches by coupling two optical parametric oscillators (OPOs) and a dual lock-in-amplifier and thus further increasing the acquisition speed [6]. More recently CRS technology co implemented with SHG and multi-photon modalities has been fitted into a compact, portable device which was successfully used in clinical imaging translation by obtaining non-invasive skin biopsies from patients suffering from psoriasis and atopic dermatitis [7]. In commercially available SRS systems, hyperspectral data is commonly acquired through continuous spectral tuning during which Stokes laser wavelength remains constant and pump laser wavelength is tuned. This process is time consuming and hence we envision that next-gen SRS optical systems will also exercise the application of multiplex SRS. In multiplex SRS, the tuneable narrowband pump laser is replaced by a broadband laser with a 32-channel lock-in-amplifier (figure 3) allowing virtually to scan through the CH region in one raster scan [8]. This speed would then almost revolutionize the capability of SRS in label-free live-cell imaging and in vivo.

In order to detect and track target molecules in biological environment, molecules of interest require chemical modifications through introduction of -C-D, -C≡C, and -C≡N moieties. These bonds vibrate at frequencies in “cell-silent” regions of the Raman spectrum and therefore can be detected at high contrast and potentially at low concentrations as there is no endogenous vibrational source within that spectral region. Ponatinib [9] and erlotinib [10], both anti-tumour drugs with an integrated alkyne bond in their structures were successfully detected with SRS in cell cultures after one hour of incubation. Entire cell components can be visualised, with SRS, through a combination of alkynes with (deoxy) ribonucleosides, amino acids, fatty acids [11]. The variety of individual Raman tags in a single cell sample exceeds the amount of fluorescent dyes significantly owing to relative narrow spectral bands intrinsic to nitrile and alkyne vibrations. Tuning the spectral resonance of Raman tags has been achieved through the introduction of changes in the chemical structure of the tag. By varying the amount of xanthene rings, substituting single atom with an alternative element and through isotope editing, Wei et. al. introduced a pallet of 20 polyyne dyes for Raman multimodal imaging [12]. Isotope editing can be further picked-up in different imaging techniques, such as PET (18F) and MRI (19F) imaging.